TE-1822 Web

IAEA-TECDOC-1822
IAEA-TECDOC-1822
IAEA TECDOC SERIES
Development of a Reference Database for Particle Induced Gamma Ray Emission (PIGE) Spectroscopy
IAEA-TECDOC-1822
Development of a Reference
Database for Particle Induced
Gamma Ray Emission
(PIGE) Spectroscopy
International Atomic Energy Agency

Vienna
ISBN 978–92–0–106317–5
ISSN 1011–4289
@
pc7209_TE-1822_cover.indd 1-3 2017-08-29 17:40:40
DEVELOPMENT OF
A REFERENCE DATABASE
FOR PARTICLE INDUCED GAMMA RAY
EMISSION (PIGE) SPECTROSCOPY
The following States are Members of the International Atomic Energy Agency:
AFGHANISTAN GEORGIA OMAN

ALBANIA GERMANY PAKISTAN
ALGERIA GHANA PALAU
ANGOLA GREECE PANAMA
ANTIGUA AND BARBUDA GUATEMALA PAPUA NEW GUINEA
ARGENTINA GUYANA PARAGUAY
ARMENIA HAITI PERU
AUSTRALIA HOLY SEE PHILIPPINES
AUSTRIA HONDURAS POLAND
AZERBAIJAN HUNGARY PORTUGAL
BAHAMAS ICELAND QATAR
BAHRAIN INDIA REPUBLIC OF MOLDOVA
BANGLADESH INDONESIA ROMANIA
BARBADOS IRAN, ISLAMIC REPUBLIC OF RUSSIAN FEDERATION
BELARUS IRAQ RWANDA
BELGIUM IRELAND SAN MARINO
BELIZE ISRAEL SAUDI ARABIA
BENIN ITALY SENEGAL
BOLIVIA, PLURINATIONAL JAMAICA SERBIA
STATE OF JAPAN SEYCHELLES
BOSNIA AND HERZEGOVINA JORDAN SIERRA LEONE
BOTSWANA KAZAKHSTAN SINGAPORE
BRAZIL KENYA SLOVAKIA
BRUNEI DARUSSALAM KOREA, REPUBLIC OF SLOVENIA
BULGARIA KUWAIT SOUTH AFRICA
BURKINA FASO KYRGYZSTAN SPAIN
BURUNDI LAO PEOPLE’S DEMOCRATIC SRI LANKA
CAMBODIA REPUBLIC SUDAN
CAMEROON LATVIA SWAZILAND
CANADA LEBANON SWEDEN
CENTRAL AFRICAN LESOTHO SWITZERLAND
REPUBLIC LIBERIA SYRIAN ARAB REPUBLIC
CHAD LIBYA TAJIKISTAN
CHILE LIECHTENSTEIN THAILAND
CHINA LITHUANIA THE FORMER YUGOSLAV
COLOMBIA LUXEMBOURG REPUBLIC OF MACEDONIA
CONGO MADAGASCAR TOGO
COSTA RICA MALAWI TRINIDAD AND TOBAGO
CÔTE D’IVOIRE MALAYSIA TUNISIA
CROATIA MALI TURKEY
CUBA MALTA TURKMENISTAN
CYPRUS MARSHALL ISLANDS UGANDA
CZECH REPUBLIC MAURITANIA UKRAINE
DEMOCRATIC REPUBLIC MAURITIUS UNITED ARAB EMIRATES
OF THE CONGO MEXICO UNITED KINGDOM OF
DENMARK MONACO GREAT BRITAIN AND
DJIBOUTI MONGOLIA NORTHERN IRELAND
DOMINICA MONTENEGRO UNITED REPUBLIC
DOMINICAN REPUBLIC MOROCCO OF TANZANIA
ECUADOR MOZAMBIQUE UNITED STATES OF AMERICA
EGYPT MYANMAR URUGUAY
EL SALVADOR NAMIBIA UZBEKISTAN
ERITREA NEPAL VANUATU
ESTONIA NETHERLANDS VENEZUELA, BOLIVARIAN
ETHIOPIA NEW ZEALAND REPUBLIC OF
FIJI NICARAGUA VIET NAM
FINLAND NIGER YEMEN
FRANCE NIGERIA ZAMBIA
GABON NORWAY ZIMBABWE
The Agency’s Statute was approved on 23 October 1956 by the Conference on the Statute of the
IAEA held at United Nations Headquarters, New York; it entered into force on 29 July 1957.
The Headquarters of the Agency are situated in Vienna. Its principal objective is “to accelerate and enlarge
the contribution of atomic energy to peace, health and prosperity throughout the world’’.
IAEA-TECDOC-1822
DEVELOPMENT OF
A REFERENCE DATABASE
FOR PARTICLE INDUCED GAMMA RAY
EMISSION (PIGE) SPECTROSCOPY
INTERNATIONAL ATOMIC ENERGY AGENCY

VIENNA, 2017
COPYRIGHT NOTICE
All IAEA scientific and technical publications are protected by the terms of
the Universal Copyright Convention as adopted in 1952 (Berne) and as revised
in 1972 (Paris). The copyright has since been extended by the World Intellectual
Property Organization (Geneva) to include electronic and virtual intellectual
property. Permission to use whole or parts of texts contained in IAEA publications
in printed or electronic form must be obtained and is usually subject to royalty
agreements. Proposals for non-commercial reproductions and translations are
welcomed and considered on a case-by-case basis. Enquiries should be addressed
to the IAEA Publishing Section at:
Marketing and Sales Unit, Publishing Section

Vienna International Centre
PO Box 100
1400 Vienna, Austria
fax: +43 1 2600 29302
tel.: +43 1 2600 22417
email: sales.publications@iaea.org
http://www.iaea.org/books
For further information on this publication, please contact:
Nuclear Data Section

Vienna International Centre
PO Box 100
1400 Vienna, Austria
Email: Official.Mail@iaea.org
© IAEA, 2017
Printed by the IAEA in Austria
September 2017
IAEA Library Cataloguing in Publication Data
Names: International Atomic Energy Agency.

Title: Development of a reference database for particle induced gamma ray emission (PIGE)
spectroscopy / International Atomic Energy Agency.
Description: Vienna : International Atomic Energy Agency, 2017. | Series: IAEA TECDOC
series, ISSN 1011–4289 ; no. 1822 | Includes bibliographical references.
Identifiers: IAEAL 17-01109 | ISBN 978–92–0–106317–5 (paperback : alk. paper)
Subjects: LCSH: Nuclear reactions. | Ion bombardment. | Cross sections (Nuclear physics). |
Gamma ray spectrometry.
FOREWORD
Providing nuclear data for use in ion beam analysis (IBA) has been a significant part of the
IAEA mission to provide its Member States with reliable nuclear data for energy and
non-energy applications. The modern analytical techniques that make use of ion beams are
used in a broad range of applications, including microelectronics, cultural heritage studies,
forensics, biology and materials sciences. These techniques are also employed in the
development of fusion technologies to study and test structural materials.
In response to the growing need for a systematic and traceable compilation of nuclear reaction
data for IBA purposes, the IAEA held a coordinated research project (CRP) from 2005 to
2010, entitled Development of a Reference Database for Ion Beam Analysis. The CRP
produced the Ion Beam Analysis Nuclear Data Library (IBANDL), which has since served the
IBA community by making readily available all existing and published experimental data on
charged particle elastic scattering and nuclear reaction cross-sections.
However, the goal of making IBANDL a comprehensive database was not fully reached,
since cross-section data for nuclear reactions that produce gamma rays were not taken into
consideration. At that time, particle induced gamma ray emission (PIGE) spectroscopy, a
powerful analytical technique that employs ion beams, had not been developed to its full
potential, mainly owing to the lack of consistent and reliable cross-section data.
Given these circumstances, the IAEA initiated a second CRP, entitled Development of a
Reference Database for Particle-Induced Gamma-ray Emission (PIGE) Spectroscopy. The
project produced new measurements, to resolve existing discrepancies and to fill in the gaps
in existing data. It also determined the accuracy and precision limits of the PIGE analysis
technique and set the groundwork for the full development of the standardless PIGE by
making available a PIGE simulation computer program. IBANDL now contains the majority
of the existing experimental cross-section data for IBA and, going forward, will require only
regular updates in the form of routine maintenance.
The IAEA wishes to thank the participants of the CRP for their contributions both to the
extension of IBANDL for PIGE analysis and to the present publication. The IAEA is grateful
to M. Stanley (United States of America), who contributed significantly to the preparation of
the publication. The IAEA officer responsible for this publication was P. Dimitriou of the
Division of Physical and Chemical Sciences.
EDITORIAL NOTE
This publication has been prepared from the original material as submitted by the contributors and has not been edited by the editorial
staff of the IAEA. The views expressed remain the responsibility of the contributors and do not necessarily represent the views of the
IAEA or its Member States.
Neither the IAEA nor its Member States assume any responsibility for consequences which may arise from the use of this publication.
This publication does not address questions of responsibility, legal or otherwise, for acts or omissions on the part of any person.
The use of particular designations of countries or territories does not imply any judgement by the publisher, the IAEA, as to the legal
status of such countries or territories, of their authorities and institutions or of the delimitation of their boundaries.
The mention of names of specific companies or products (whether or not indicated as registered) does not imply any intention to
infringe proprietary rights, nor should it be construed as an endorsement or recommendation on the part of the IAEA.
The IAEA has no responsibility for the persistence or accuracy of URLs for external or third party Internet web sites referred to in this
publication and does not guarantee that any content on such web sites is, or will remain, accurate or appropriate.
CONTENTS
1. INTRODUCTION ......................................................................................................... 1
1.1. BACKGROUND ............................................................................................. 1
1.2. OBJECTIVE.................................................................................................... 2
1.3. SCOPE ............................................................................................................ 2
1.3.1. Compilation of existing data ................................................................. 2
1.3.2. New measurements............................................................................... 2
1.3.3. Co-ordinated Measurements ................................................................. 3
1.3.4. Evaluations of PIGE cross-sections ...................................................... 3
1.3.5. PIGE Analysis Code ............................................................................. 3
1.3.6. Updated IBANDL ................................................................................ 3
1.4. STRUCTURE .................................................................................................. 5
REFERENCES TO SECTION 1............................................................................................. 6

2. PIGE CONCEPTS AND IMPLEMENTATION ............................................................ 7
2.1. PIGE BULK ANALYSIS ................................................................................ 7
2.1.1. PIGE bulk analysis ............................................................................... 7
2.1.2. Production and acquisition of gamma-ray spectra ................................. 7
2.1.3. Determination of isotopic/elemental concentrations ............................ 13
2.1.4. The standard-based method ................................................................ 14
2.1.5. The standardless method..................................................................... 14
2.2. PIGE DEPTH PROFILING WITH NUCLEAR RESONANCES ................... 16
2.3. A TYPICAL MINIMAL PIGE FACILITY .................................................... 20
2.3.1. Particle accelerator ............................................................................. 21
2.3.2. Charge collection and measuring system ............................................ 22
2.3.3. Gamma-ray detector ........................................................................... 23
2.3.4. Example of a minimal PIGE set-up..................................................... 25
2.3.5. Example of an in-vacuum multi detector system ................................. 25
2.4. EXAMPLES OF APPLICATIONS IN NA BULK ANALYSIS ..................... 26
2.4.1. Identification of lapis-lazuli in paint layers ......................................... 26
2.4.2. Quantification of Na in ancient glasses ............................................... 28
2.5. HYDROGEN DEPTH PROFILING .............................................................. 29
REFERENCES TO SECTION 2........................................................................................... 33

3. METHODOLOGY OF CROSS-SECTION MEASUREMENTS ................................. 37
3.1. ENERGY CALIBRATION OF THE ACCELERATOR ................................ 37
3.1.1. Narrow resonances ............................................................................. 37
3.1.2. Neutron producing threshold reactions ............................................... 38
3.1.3. Elastic scattering ................................................................................ 39
3.2. EFFICIENCY CALIBRATION OF THE DETECTOR.................................. 39
3.3. TARGET PREPARATION............................................................................ 40
3.4. CHARGE COLLECTION ............................................................................. 41
3.5. CROSS-SECTION DETERMINATION ........................................................ 42
3.6. INTERLABORATORY EXERCISE: MEASUREMENT OF THE
CROSS-SECTION OF THE 27AL(P,P′T)27AL REACTION .......................... 43
3.6.1. Implementation .................................................................................. 43
3.6.2. Experiment ......................................................................................... 45
3.6.3. Results and discussion ........................................................................ 50
REFERENCES TO SECTION 3........................................................................................... 57
4. MEASUREMENTS OF PARTICLE-INDUCED GAMMAR-RAY EMISSION
(PIGE) CROSS-SECTIONS ........................................................................................ 61
4.1. LITHIUM ...................................................................................................... 62
4.1.1. Gamma-ray production cross-cections of lithium from proton-
induced nuclear reactions ................................................................... 62
4.1.2. Gamma-ray production cross-sections of lithium from deuteron-
4.2. BERYLLIUM ................................................................................................ 75
4.2.1. Gamma-ray production cross-sections of beryllium from proton-
4.2.2. Gamma-ray production cross-sections of beryllium from
deuteron-induced nuclear reactions ..................................................... 77
4.2.3. Gamma-ray production cross-sections of beryllium from alpha-
4.3. BORON ......................................................................................................... 79
4.3.1. Gamma-ray production cross-sections of boron from proton-
4.3.2. Gamma-ray production cross-sections of boron from deuteron-
4.4. CARBON ...................................................................................................... 95
4.4.1. Gamma-ray production cross-sections of carbon from proton-
4.4.2. Gamma-ray production cross-sections of carbon from deuteron-
4.5. NITROGEN................................................................................................. 103
4.5.1. Gamma-ray production cross-sections of nitrogen from proton-
induced nuclear reactions ................................................................. 103
4.5.2. Gamma-ray production cross-sections of nitrogen from deuteron-
4.6. FLUORINE ................................................................................................. 117
4.6.1. Gamma-ray production cross-sections of fluorine from proton-
4.6.2. Thick target yields ............................................................................ 128
4.7. SODIUM ..................................................................................................... 132
4.7.1. Gamma-ray production cross-sections of sodium from proton-
4.8. MAGNESIUM ............................................................................................ 144
4.8.1. Gamma-ray production cross-sections of magnesium from proton-
4.9. ALUMINUM ............................................................................................... 154
4.9.1. Gamma-ray production cross-sections of aluminum from proton-
4.10. SILICON ..................................................................................................... 171
4.10.1. Gamma-ray production cross-sections of silicon from proton-
induced nuclear reactions................................................................. 171
4.10.3. Gamma-ray production cross-sections of silicon from deuteron-
4.11. PHOSPHORUS ........................................................................................... 182
4.11.1. Gamma-ray production cross-sections of phosphorus from proton-
4.12. SULPHUR ................................................................................................... 186
4.12.1. Gamma-ray production cross-sections of sulphur from proton-
4.12.2. Gamma-ray production cross-sections of sulphur from deuteron-
4.13. TITANIUM ................................................................................................. 191
4.13.1. Gamma-ray production cross-sections of titanium from proton-
4.14. CHROMIUM............................................................................................... 196
4.14.1. Gamma-ray production cross-sections of chromium from proton-
REFERENCES TO SECTION 4......................................................................................... 199

5. THE EMITTED RADIATION YIELD ANALYSIS (ERYA) CODE FOR PIGE
ANALYSIS ............................................................................................................... 207
5.1. INTRODUCTION .............................................................................................. 207
5.2. ERYA .......................................................................................................... 207
5.2.1. ERYA-bulk ...................................................................................... 207
5.2.2. ERYA-profiling................................................................................ 213
5.3. USER FEEDBACK AND FUTURE DEVELOPMENTS ............................ 220
REFERENCES TO SECTION 5......................................................................................... 221

6. CONCLUSIONS ....................................................................................................... 222
APPENDIX ........................................................................................................................ 224
LIST OF CONTRIBUTORS TO DRAFTING AND REVIEW........................................... 231
1. INTRODUCTION
1.1. BACKGROUND
Ion beam analysis (IBA) is an important group of analytical techniques that is widely used in
applications ranging from analysis of fission reactor material to biomedicine, environment,
cultural heritage, and more recently, fusion reactor technologies. The technique uses energetic
ion beams to probe the surface of materials in order to determine composition. All IBA
methods are non-destructive, highly sensitive and allow the detection of elements in depths
ranging up to several hundreds of micrometers.
Among the various existing types of IBA techniques, Particle-Induced Gamma-ray Emission
(PIGE) spectroscopy, based on gamma-producing nuclear reactions, is one of the most
powerful. The PIGE technique involves measuring the energy spectra of the prompt gamma-
rays that are emitted as a result of accelerated charged particles hitting the target matter.
Based on these measurements, the composition and structural details of the surface region of
the target matter can be determined. This technique has been used regularly since the early
1960's, when it was discovered that it has the potential for performing depth profiling of the
surface region of materials with better resolution than other IBA techniques. This unique
characteristic of PIGE has proven particularly useful in the development of hydrogen depth
profiling: a technique which is now widely used.
As with all IBA methods, including Rutherford backscattering (RBS), elastic backscattering
(EBS) and nuclear reactions analysis (NRA), PIGE relies largely on the availability of
differential cross-section data. Our poor knowledge of cross-sections for nuclear reactions
using charged particles accelerated to only a few megaelectron-volts of energy has
significantly hindered progress in the implementation of IBA techniques in general. To
address the IBA data needs, a Coordinated Research Project (CRP) on the Development of a
Reference Database for Ion Beam Analysis from 2005 to 2010 was initiated [1.1]. Over the
course of these five years, the CRP produced a comprehensive and reliable database of cross-
section data and evaluated cross-sections that are primarily relevant for eEBS and NRA
techniques. The experimental cross-section data are available on the Ion Beam Analysis
Nuclear Data Library (IBANDL) [1.2] and the evaluated cross-sections are available through
the online calculator SigmaCalc [1.3].
Although the CRP from 2005 to 2010 met the data needs for EBS and NRA techniques, the
data sets available for PIGE remained minimal. There existed a considerable amount of
published data on PIGE within the nuclear physics literature, but no current and
comprehensive compilation dedicated to IBA purposes. Furthermore, although some PIGE
cross-section data had already been uploaded to IBANDL by members of the IBA
community, a preliminary survey of the data revealed numerous discrepancies. It was deemed
necessary to resolve these discrepancies and add more data sets to those already available, to
allow the unfolding of the advantages of the PIGE technique without having to resort to using
known ‘standards’ and further enhance the implementation of ‘standardless’ PIGE.
In response, a consultant's meeting to assess the situation and propose actions for improving
the quality and availability of nuclear cross-section PIGE data was organized. The meeting
was held in 2010 [1.4] and resulted in the recommendation of a CRP.
1
1.2. OBJECTIVE
The IAEA CRP on ‘Development of a Reference Database for Particle-Induced Gamma-ray

Emission (PIGE) spectroscopy was subsequently established in 2011 with the overall
objective of producing updated and reliable nuclear data for PIGE analysis and making these
data freely available to the user community.
1.3. SCOPE
To achieve the overall objective of the CRP, the following areas of work with corresponding
deliverables and reports were defined:
1.3.1. Compilation of existing data
1) Exhaustive survey of the existing literature to find all the data regarding nuclear reactions
with the nuclides deemed the most important for studies using the PIGE technique, i.e.,
Li, B, Be, 12C, N, F, Na, Mg, Al, Si, P, Cl, S, Ti and Cr.
2) Identify discrepancies between measurements and which measurements were missing or

needed to be repeated.
3) Compilation of retrieved data in IBANDL.
4) Deliverables: (i) List of compiled data complete with references and (ii) Updated
IBANDL.
1.3.2. New measurements
1) New measurements of differential cross-sections and thick target yields of gamma-ray

producing reactions listed in Table A.1 of the Appendix at different angles in order to
obtain angular distributions of the emitted gamma-rays to establish the isotropy or
anisotropy of the reactions studied.
2) Deliverables: (i) cross-sections and thick-target yields data files, (ii) Updated IBANDL,
(iii) Publications of new measurements in peer-reviewed scientific journals, (iv) Reporting
of progress in Research Coordination Meeting (RCM) reports and scientific journals [1.5-
1.8].
2
1.3.3. Co-ordinated Measurements
Perform an inter-laboratory exercise to calibrate equipment and co-ordinate measurements:
1) Measurement of the excitation function of the 27Al(p,pγ1–0)27Al reaction, which is known

to emit a gamma-ray with an energy (Eγ) of 844 keV, at incident energies ranging from
2.5 to 3 MeV at intervals of 10 keV. Since the cross-section of this reaction is known to be
isotropic due to spin considerations, the angles of measurements do not matter.
2) Measurements of thick target yields using the same targets that are routinely employed for
accelerator energy calibration. The aim was to reproduce the resonance of the
27
Al(p,γ)28Si reaction at the proton energy of 0.991 MeV by measuring the 1.779 MeV
gamma-ray that this reaction is known to produce.
1.3.4. Evaluations of PIGE cross-sections
Evaluation of PIGE cross-sections using the same methodology that was applied to the
evaluation of the charged-particle cross-sections in SigmaCalc [1.3].
Deliverables: (i) Evaluated data files for the following reactions 27Al(p,p′γ)27Al,
23
Na(p,p′γ)23Na and 52Cr(p,γ)53Mn in IBANDL, (ii) Report of evaluation [1.9, 1.10].
1.3.5. PIGE Analysis Code
Development of a PIGE analysis code to calculate all the integral quantities needed in PIGE
analyses of bulk samples and also in depth profiling. In particular, the code must be able to
calculate the material composition in the case of bulk analysis and the layer composition in
the case of depth profiling by using the differential cross-section data and stopping power
ranges that are available in the corresponding databases.
Deliverable: PIGE analysis code (ERYA) available on the web including a detailed manual
[1.11].
1.4. IBANDL
The final end-product of the CRP is an updated and complete IBANDL. All the new cross-
section data and thick target yields compiled and measured during the course of the CRP have
been made available in IBANDL. The number of published articles containing PIGE data
found during the CRP (2011–2015) on the nuclides listed earlier in is illustrated in Fig. 1.1,
while the distribution of new measurements over the nuclides is shown in Fig. 1.2.
3
FIG. 1.1. Number of articles with published PIGE data for particular nuclides that were found during
the first step of the CRP (2011–2015). All data are available in IBANDL.
FIG. 1.2. Number of measurements for various nuclides performed during the CRP (2011–2015). All
measurements are available in IBANDL
The tremendous expansion of PIGE data sets included in IBANDL is illustrated in Fig. 1.3.
Although the majority of the expansion of data occurred during the CRP, some data sets
continued to be uploaded to IBANDL continuing through 2016. To date over 600 PIGE data
sets have been uploaded.
4
FIG. 1.3. Total number of PIGE data sets in IBANDL before the CRP (2008) in comparison to each
consecutive year during the course of the CRP (2013–2015).
In the past decade, the concerted efforts made within the two CRPs, has led to a huge increase
of EBS, NRA and PIGE data sets in IBANDL which is illustrated in Fig. 1.4. IBANDL is
currently the only online database dedicated to experimental nuclear cross-section data for
IBA applications that is comprehensive and continuously up-to-date.
FIG. 1.4. Total number of EBS, NRA and PIGE data sets in the IBANDL database before either CRP
(2008) compared to the end of both CRPs (2015) and currently (2016).
1.5. STRUCTURE
Section 2 of the report provides an overview of the concepts and implementation of the PIGE
analysis technique, while Section 3 outlines the methodology applied in cross-section and
thick-target yield measurements. The data measured during the CRP are presented and
assessed in Section 4, and the PIGE analysis code (ERYA) is described in Section 5. Some
conclusions are given in Section 6.
5
REFERENCES TO SECTION 1
[1.1] INTERNATIONAL ATOMIC ENERGY AGENCY, Development of a Reference

Database for Ion Beam Analysis, IAEA-TECDOC-1780, IAEA, Vienna (2015).
[1.2] Ion Beam Analysis Nuclear Data Library (IBANDL), http://www-
nds.iaea.org/ibandl/
[1.3] SigmaCalc Online Calculator, http://sigmacalc.iate.obninsk.ru/
[1.4] ABRIOLA, D., PEDRO DE JESUS, A., Assessment of Nuclear Data Needs for
Particle Induced Gamma Ray Emission (PIGE) (Proc. Consultants’ Mtg. Vienna,
2010), INDC(NDS)--0568, IAEA, Vienna (2010).
[1.5] ABRIOLA, D., PEDRO DE JESUS, A., Development of a Reference Database for
Particle-induced Gamma Ray Emission (PIGE) Spectroscopy (Proc. 1st Res.
Coordination mtg. Vienna, 2011), INDC(NDS)-0589, IAEA, Vienna (2011).
[1.6] ABRIOLA, D., DIMITRIOU, P., PEDRO DE JESUS, A., Development of a
Reference Database for Particle-induced Gamma Ray Emission (PIGE) Spectroscopy
(Proc. 2nd Res. Coordination mtg. Vienna, 2012), INDC(NDS)-0625, IAEA, Vienna
(2012).
[1.7] DIMITRIOU, P., PEDRO DE JESUS, A., Development of a Reference Database for
Particle-induced Gamma Ray Emission (PIGE) Spectroscopy (Proc. 3rd Res.
Coordination mtg. Vienna, 2014), INDC(NDS)-0664, IAEA, Vienna (2014).
[1.8] DIMITRIOU, P., BECKER, H.-W., BOGDANOVIĆ-RADOVIĆ, I., CHIARI, M.,
GONCHAROV, A., JESUS, A.P., KAKUEE, O., KISS, A.Z., LAGOYANNIS, A.,
RÄISÄNEN, J., STRIVAY, D., ZUCCHIATTI, A., Development of a reference
database for particle-induced gamma-ray emission spectroscopy, Nucl. Instrum.
Methods Phys. Res. B 317 (2016) 33–36.
[1.9] DIMITRIOU, P. (Ed.), Developments in the Ion Beam Analysis Nuclear Data
Library (IBANDL), INDC(NDS)-0660, IAEA, Vienna (2014).
[1.10] GURBICH, A.F., Evaluation of cross-sections for particle induced gamma-ray
emission (PIGE) spectroscopy, Nucl. Instrum. Methods Phys. Res. B 331 (2014) 31–
33.
[1.11] FONSECA, M., JESUS, A.P., LUÍS, H., MATEUS, R., CRUZ, J., GASQUES, L.,
GALAVIZ, D., RIBEIRO, J.P., PIGE analysis of magnesium and beryllium, Nucl.
Instrum. Methods Phys. Res. B 268 11–12 (2010) 1806–1808.
[1.12] VICKRIDGE, I.C., Proposed ASCII Format for Communication of Reaction Cross-
sections in the IBA Community, DSIR Physical Sciences Report 33, Geological and
Nuclear Sciences Ltd, Lower Hutt, (2004).
[1.13] MAYER, M., SIMNRA: Computer simulation of RBS, ERDA and NRA, Max-
Planck-Institut für Plasmaphysik, Garching, http://home.mpcdf.mpg.de/~mam/
[1.14] BARRADAS, N.P., JEYNES, C., Advanced physics and algorithms in the IBA
DataFurnace, Nucl. Instrum. Methods Phys. Res. B 266 8 (2008) 1875–1879.
[1.15] NDF - General purpose code for data analysis of Ion Beam Analysis Data,
http://www.itn.pt/facilities/lfi/ndf/uk_lfi_ndf.htm
6
2. PIGE CONCEPTS AND IMPLEMENTATION
2.1. PIGE BULK ANALYSIS
The implementation of the PIGE technique began in the 1960s when NaI (Tl) detectors were
still used [2.1–2.4]. However, the full multi-elemental analytical potential of the technique
was only revealed in the 1970s by taking advantage of the high resolution of Ge(HP) detectors
[2.5, 2.6]. In 1980, Borderie et al. [2.5] introduced the methodology for homogeneous thick
samples, i.e. PIGE bulk analysis. Boni et al. adapted that methodology for thin target analysis
and measured several cross-sections of gamma-emitting reactions [2.7–2.9]. Other efforts to
measure and compile useful information, such as cross-sections and thick target yields
followed [2.10–2.16]. In the last decade, a standardless methodology for bulk analysis, based
on the integration of the pertinent nuclear reaction along the depth of the sample, was
developed [2.17–2.22]. Although deuterons have been explored as analyzing beams, most of
the work done so far is based on protons.
2.1.1. PIGE bulk analysis
When protons of a few MeV bombard a sample, for some light elements, such as Li, Be, B, F,
Na, Mg, Al, the cross-sections of a gamma-producing reaction, such as (p,αγ) or (p,p′γ) reach
the hundreds of millibarns and, consequently, a high yield of characteristic gamma-rays are
produced, which may be used to detect and quantify the corresponding isotopes (elements).
Since the referred elements are generally not detectable, or can only be quantified with a large
degree of uncertainty by PIXE (Particle-Induced X-Ray Emission), PIGE can be used as a
complementary analytical technique. The PIXE technique in this case is based on the specific
details of the experimental set-up and the employed X-ray detector, which includes variable
such as in-vacuum or external-beam set-up, detector entrance window thickness and material.
In the following subsections, the basic procedure for a typical PIGE bulk analysis is
discussed.
2.1.2. Production and acquisition of gamma-ray spectra
Instrumentation for PIGE measurements is, in principle, very similar to the one used for
PIXE, with the exception of the radiation detector, which must have a full energy peak
intrinsic efficiency that is suited for detecting radiation in the range above 100 keV. In
comparison, X-rays are in the range of a few to tens of keV. Therefore, the following
experimental conditions must be met:
(a) Accelerator (1–3 MV terminal voltage) — which together with the associated electro-
magnetic beam transport system must assure a very good definition of energy (a
resolution of 1 keV or less);
(b) Reaction chamber — with a 10-6 mbar or better vacuum produced by, preferably, free-
oil pumps, in order to avoid the deposition of carbon films on the sample. PIGE bulk
analysis can also be easily and proficiently performed in external beam or ambient
pressure set-ups, which reduces the risk to the samples in terms of charging, heating and
carbon build-up effects;
7
(c) Beam charge collection system — with implemented conditions to avoid the loss of
secondary electrons sputtered from the target, if analysis is done in vacuum, and to
measure the correct value of the beam charge;
(d) Target holder — with multi-target capacity and degrees of freedom that allow changing
between the samples and the ability to tilt them with respect to the incoming beam
direction, if necessary;
(e) Gamma-ray detector — for multi-elemental analysis, HPGe or Ge(Li) detectors with a
few keV energy resolution must be used. For standard free analysis, this detector must
be calibrated in efficiency by radioactive sources certified in activity and/or by Monte-
Carlo simulations (codes Geant, Penelope).
More details about a typical set-up are given in Section 2.3.
The definition of the beam energy is complete only after performing a thorough accelerator
energy calibration. For very stable Cockroft and Walton type accelerators, the high voltage
terminal values may be calibrated. Another choice is the calibration of the magnetic fields of
the analyzing magnets, measured by Nuclear Magnetic Resonance (NMR) probes. In either
case, nuclear reaction resonances provide the beam energy values. Several resonances have
been employed for this purpose. In Appendix 19 of the Handbook of Modern Ion Beam
Materials Analysis, a description of calibration methods and a table of resonances can be
found [2.23]. More alternatives are described in Section 3.2.
Care should be taken when using resonance energies from the literature. For strong, isolated
resonances like the one for 27Al(p,γ)28Si at 991.6 keV, the proton energy value corresponding
to half-the-height of a thick (compared to the width of the resonance) target yield curve gives
the correct value of the resonance energy. For other resonances, thin target resonances may be
preferable, but one should note that the experimental maximum of the resonance may deviate
from the tabulated resonance energy, ER, due to the target thickness. Where ∆E is the target
thickness and Γ is the natural width of the resonance: for ∆E << Γ, the maximum corresponds
to ER; for ∆E >> Γ the maximum corresponds to ER + ∆E/2; for other intermediate situations,
intermediate values between ER and ER + ∆E/2 are used.
8
10000 Na
Li
B
Al
Gamma ray yield
100
200 400 600 800 1000

E (keV)
FIG. 2.1. Gamma-ray spectrum from a mineral sample containing fluorine, boron, sodium, lithium
and aluminium. It is represented in logarithmic scale to enhance the continuum background from
Compton interactions of the radiation with the detector.
Table 2.1 summarizes some of the gamma-ray lines which may be used for PIGE analysis, for
different light isotopes (mass A < 40). These may be easily separated by high resolution
detection as illustrated in Fig. 2.1. It can also be observed that these lines are sitting on a
continuum produced by Compton interactions in the detector. Although some of the
background originates from environmental and cosmic radiation, most of it is due to prompt
gamma radiation from the beam bombarding the target or from the interaction of the beam
with collimators or other parts of the measurement set-up. Hence, the continuum background
is mostly dependent on the composition of the target and also on the specific experimental set-
up. This implies that it is not possible to define general sensitivity limits [2.24]. For example,
the 19F(p,αγ)16O reaction, which has the highest cross-section of all nuclear reactions,
produces high-energy gamma-rays (~6–7 MeV), which contribute to the continuum Compton
background. The presence of fluorine in a sample is thus a factor that limits the sensitivity of
the detection of the other elements. The best sensitivity values that can be attained are of the
order of µg/g for fluorine, lithium and sodium at energies equal to or higher than 2.4 Me
9
TABLE 2.1. PIGE REACTIONS ON LIGHT NUCLEI AND CORRESPONDING GAMMA-RAY
ENERGIES
Eγ (keV) Reaction Isotopic abundance (%)
7
429* Li(p, nγ)7Be 93
7
478 Li(p, pγ)7Li
429* 10
B(p, αγ)7Be 19
10
718 B(p, pγ)10B
11
2125 B(p, pγ)11B 81
19
110 F(p, pγ)19F 100
19
197 F(p, pγ)19F
19
1236 F(p, pγ)19F
19
1349 F(p, pγ)19F
19
1357 F(p, pγ)19F
19
1459 F(p, pγ)19F
6129 19
F(p, αγ)16O
25
390 Mg(p,pγ)25Mg 10.1
25
417 Mg(p, γ)26Al
24
452 Mg(p,γ)25Al 78.7
25
585 Mg(p,pγ)25Mg
26
844* Mg(p, γ)27Al 11.3
25
975 Mg(p,pγ)25Mg
25
990 Mg(p,pγ)25Mg
26
1014* Mg(p, γ) 27Al
24
1369* Mg(p,pγ)24Mg
10
ENERGIES (cont.)
25
1380 Mg(p,pγ)25Mg
25
1612 Mg(p,pγ)25Mg
26
1809 Mg(p,pγ)26Mg
25
1965 Mg(p,pγ)25Mg
25
390 Mg(p,pγ)25Mg 10.1
27
171 Al(p, pγ)27Al
27
844* Al(p, pγ)27Al
27
1014* Al(p, pγ)27Al 100
1369* 27
Al(p, αγ)24Mg
27
1720 Al(p, pγ)27Al
27 28
1779 Al(p, γ) Si
27
2211 Al(p, pγ)27Al
27
2734 Al(p, pγ)27Al
27
2839 Al(p,γ) 28Si
29
755 Si(p, pγ)29Si 4.71
30
1266* Si(p, γ) 31P 3.12
29
1273 Si(p, pγ)29Si
28
1779* Si(p, pγ)28Si 92.2
29
2028 Si(p, pγ)29Si
23
440 Na(p,pγ)23Na 100
23
1369 Na(p, γ)24Mg
1634 Na(p, αγ)20Ne

23
23
1636 Na(p,pγ)23Na
23
1951 Na(p,pγ)23Na
11
ENERGIES (cont.)
23
2391 Na(p,pγ)23Na
23
440 Na(p,pγ)23Na 100
23
1369 Na(p, γ)24Mg
1634 Na(p, αγ)20Ne

23
23
1636 Na(p,pγ)23Na
23
1951 Na(p,pγ)23Na
23
2391 Na(p,pγ)23Na
23
440 Na(p,pγ)23Na 100
31
1266* P(p, pγ)31P 100
1779* P(p, αγ)28Si

31
31
2230 P(p, γ)32S
31
2233 P(p, pγ)31P
31
1266* P(p, pγ)31P 100
1779* P(p, αγ)28Si

31
31
2230 P(p, γ)32S
30
2233 Si(p, γ)31P
30
2235 Si(p, pγ)30Si
* γ-rays emitted by more than one reaction
12
2.1.3. Determination of isotopic/elemental concentrations
For a thin target, where the cross-sections and stopping-power can be assumed to be constant,
the yield of gamma-rays with energy Eγ, detected at a given ion beam energy E depends on
the number of bombarding ions Q/e (Q being the collected charge and e the charge state of the
ion beam), the number of nuclei of the relevant isotope per surface unit NiA, the relevant
nuclear reaction cross-section σ (which may be dependent on the detection angle) and the
detection absolute efficiency εabs. For an isotropic gamma-line the yield may be written as:
Q Q
Y ( E ) = ε abs ( E γ ) ⋅   ⋅ σ ( E ) ⋅ N i A = ε abs ( E γ ) ⋅   ⋅ σ (E) ⋅ N i ⋅ ∆x (2.1)
e e
where N i is the number of nuclei of the relevant isotope per volume unit and ∆x is the target
linear thickness.
For a thick target, where by definition the thickness is greater than the range of the
bombarding ions, the above equation still applies, but for each thin layer that is parallel to the
target surface (for normal incidence), as is illustrated in Fig. 2.2.
For each layer of linear thickness dx or mass thickness (mass per surface unit) dΓ, a similar
equation may be written in terms of the relevant element mass fraction fm:
Q
dY(E) = ε abs (E γ ) ⋅   ⋅ σ(E) ⋅ f m ⋅ f i ⋅ N av ⋅ A −1 ⋅ dΓ (2.2)
e
where fi is the isotopic abundance, Nav is the Avogrado number and A is the atomic mass of
the relevant element.
E0 E 0 Energy
Beam
dx
0 x R Depth
FIG. 2.2. Division of a thick target in layers and relation between depth and energy dimensions.
The thick-target yield is the result of integrating the above equation over the range of the
incident ions in the target, R:
Q R
Y(E) = ε abs (E γ ) ⋅   ⋅ f m ⋅ f i ⋅ N av ⋅ A −1 ⋅ ∫ σ( Γ ) dΓ (2.3)
e 0
13
or
Q E0
Y ( E ) = ε abs ( E γ ) ⋅   ⋅ f m ⋅ f i ⋅ N av ⋅ A −1 ⋅ ∫ σ ( E ) S m ( E ) dE , (2.4)
e 0
where E0 is the incident energy of the beam and Sm(E) represents the mass stopping power
expressed in energy per areal mass units, which describes the influence of the sample
composition over the incident beam. This is a similar expression to the one used for PIXE,
except for the presence of the isotopic abundance term, which is needed only for PIGE, and
the absence of the term for the transmission of radiation through the target, which is only
needed for PIXE. The self-absorption effect for gamma-rays produced by proton
bombardment of a thick-target sample is indeed negligible.
There are currently two approaches used to determine the mass fractions of the elements in a
sample: the standard-based method and the standardless method.
2.1.4. The standard-based method
If one assumes that, in a first approximation, the stopping power is not energy dependent, the
integration in Eq. 2.4 is performed over the cross-section only and is therefore independent of
the target composition. This first approximation has been used for PIGE analysis based on
comparison with a standard sample. When irradiating the sample and a standard under the
same experimental conditions, the ratio of the mass fraction fm between an individual light
element in sample i and a standard st is given by:
f mi S mi ( E1 / 2 ) Y i
= ⋅ , (2.5)
f mst S mst ( E1 / 2 ) Y st
where Sm and Y are the corresponding mass stopping powers and yields of detected gamma-
rays and E1/2 is the beam energy for which the gamma-ray yield is half of the one obtained at
incident beam energy. A similar expression is valid for atomic fractions, if atomic stopping
powers (in energy per areal atomic density) are used.
This standard-based method has some drawbacks:
(i) It depends on the availability of standards, which should be as similar in composition to

the samples as possible (see below);
(ii) The exact major compositions of the standard and the sample must be known, in order
to calculate the stopping powers with low uncertainty;
(iii) The half energy E1/2 must be known. If the major compositions of the sample and
standard are not similar, E1/2 for the sample is not equal to E1/2 for the standard. Thick-
target yields obtained before and during this CRP, now available in IBANDL, may be
helpful.
2.1.5. The standardless method
In order to avoid standards, one may use Eq. 2.4 and perform the calculations. Similar to
PIXE, the integral in Eq. 2.4 can be calculated. It is, in principle, easier than for PIXE,
because the radiation transmission for PIGE in the sample is constant and equal to one (it is a
self-absorption free technique), but for many years it seemed to be more difficult, due to the
cross-section function (or excitation function).
14
For PIXE, the cross-section is a smooth function of the beam energy and it may be
determined from theoretical calculations (e.g. Energy-Loss Coulomb-Repulsion Perturbed-
Stationary State Relativistic Theory (CPSSR) for inner shell ionization) or from polynomial
fits to experimental data. For PIGE, the nuclear reaction excitation functions may have several
thin and broad resonances, which hinder this analytical approach. However, if cross-sections
are available in numerical form, with resonances defined in detail, the energy steps used for
the cross-section measurements may be used to define the layers that are necessary for the
subdivision of the target in thin layers and, furthermore, the integral may be calculated as a
sum of terms corresponding to these layers (Fig. 2.2):
∫
E0

Ek +1
 k
[ ]
σ(E)/ε(E) dE =∑ (1 S mk ) ⋅ ∫ σ(E) dE  = ∑ (1 S mk ) ⋅ σ(E k,k +1 ) ⋅ ( E k +1 − E k ) ⋅ , (2.6)
k 
0 Ek
where the cross-section for each layer k is the mean between σk and σk + 1, i.e. the measured
cross-sections at energies Ek and Ek + 1. As the integrated cross-section rises quickly with the
proton energy, it is not necessary to go down to zero energy, but only to the minimum energy
used for the cross-section measurement, Emin.
When performing PIGE analysis using the standardless method based on Eq. 2.4, with the
integral calculated by Eq. 2.6, one has to be aware of the value of the Emin in order to choose a
beam energy E0 for which the yield is much larger than Ymin. If this is not possible, Ymin
cannot be neglected and should be measured for the sample to be analysed and added up to
the calculated yield. When choosing the incident energy, it is important to avoid the
resonances. The choice of bombarding energy between resonances reduces the effect in the
final results of the uncertainty and/or deviations of the ion beam energy.
The propagation of the uncertainties of the cross-section values, the absolute value of the
collected charge, the detector efficiency and the stopping powers leads to an uncertainty of the
determined mass fractions which may be larger than 20%. However, this may be reduced if
calibrations are performed (as is typically done for standardless PIXE). The calibration
consists of the following procedure.
For each light element to be analysed, a comparison is made between the measured gamma-
ray yield and the calculated gamma-ray yield for either (i) a pellet made of a pure inorganic
compound that contains the element to be analysed or (ii) a standard reference sample. The
ratio of the experimental to calculated yield is then used as an external calibration factor (or
correction factor) in future analyses of the given element, provided the same experimental
conditions (e.g. ion beam, detector type and detector geometry) are used and also that the
same excitation function is used in the calculations. In order to confirm the stability of the
experimental conditions (mainly of the charge measurement), one might use a lithium or a
boron standard sample (the cross-sections related to 7Li(p,p′γ)7Li and 10B(p,p′γ)10B reactions
are smooth) and measure its yield before each analytical running time. Repeatibility and
reproductibility tests should be performed to estimate the uncertainties to be assigned to the
analytical procedure. By applying a careful calibration procedure in conjunction with a good
knowledge of the major elements’ composition of the samples to be analysed, the
uncertainties may be lowered down to 5–7%, which is currently the average uncertainty of the
stopping powers. In other words, the analytical procedure is dominated by the uncertainties
of the stopping powers.
15
This method is also valid for semi-thick targets, meaning targets inducing a proton energy
loss that is not negligible, but with a thickness smaller than the range of protons in the target.
In this case, instead of extending the sum in Eq. 2.6 to Emin, one sums down to Ef, the energy
with which protons emerge from the target. For a thin target of mass thickness equal to ∆Γ, an
expression equivalent to Eq. 2.4 applies:
Q
Y(E0 ) = ε abs (E γ ) ⋅   ⋅ σ(E 0 ) ⋅ f m ⋅ f i ⋅ N av ⋅ A −1 ⋅ ∆Γ , (2.7)
e
where E0 is the incident proton energy.
When dealing with semi-thick or thin samples, attention must be paid to the position of the
sample with respect to the beam. The angle θ, which is subtended by the normal both to the
sample surface and to the beam ,will increase the sample thickness by 1/cosθ (in comparison
with the normal position). The uncertainty in this angle will propagate into the uncertainty of
the final mass fraction results.
A code that is developed to calculate the mass fraction from the experimental yields may also
include a fitting procedure to optimize the sample composition (major and minor
elements/isotopes) by modifying the first guess introduced by the user. This can be done
simultaneously for all light elements present in the sample (see Section 5 for more details
about the ERYA code).
Hence, the standardless method, where standard samples are not required for routine analysis,
also has the possibility to adjust and improve the major composition of the sample, which is
often not completely known.
This method may also be applied to multi-layered samples. If the element to be analysed is
only present in one layer, E0 is simply replaced by the proton energy at the entrance of that
layer. However, if the element is present in more than one layer, the information obtained
from the code corresponds to the averaged concentration.
2.2. PIGE DEPTH PROFILING WITH NUCLEAR RESONANCES
As described above in Section 2.1, for the analysis of bulk PIGE measurements, the number
of emitted gamma-rays is proportional to the integral of the cross-section from zero to the
incident beam energy (Eq. 2.4). In cases where layers are to be investigated, this integral has
to be calculated from the energy given by the energy loss at the end of the layer up to the
energy of the beam at the entrance of the layer. Since the emitted gamma-ray carries no
information about the depth from which it originated from, the method is not depth sensitive.
However, since the cross-section for the reaction might vary strongly with incident energy,
the sensitivity to the reaction for a given isotope can also vary with depth. By exploiting the
energy dependence of the cross-section, it is possible to obtain moderate depth information
when a sample is investigated with different beam energies [2.25]. A more refined method,
called depth profiling, is described in the following.
Depth profiling with resonances in particle-induced gamma-ray emitting reactions makes use
of those resonances in the reaction cross-section where the cross-section is significantly
stronger than in the (non-resonant) region above and below the resonance energy. In almost
all cases of PIGE depth profiling, the incident beam is a proton beam, but for the depth
16
profiling of hydrogen, heavier beams such as 15N, 18O or 19F beams are used in the inverse
kinematics. In an ideal case, a resonance should be very strong and narrow, so that the
contributions to the reaction are limited to a narrow energy region.
Depth profiling is performed by measuring the gamma-yield of the reaction as a function of

the beam energy. When starting with energies near the resonance energy, the concentration of
an isotope is probed only at the surface of the sample, since the beam loses energy while
penetrating into the sample due to the electronic stopping and therefore the corresponding
cross-section lies in the region of low cross-section beyond the resonance. By increasing the
beam energy sequentially above the resonance energy, the beam energy will reach the
resonance energy at a certain depth, after losing energy due to electronic stopping. In visual
terms, the resonance window, i.e. the window of energy where the crossing of the ion beam
through the sample is most sensitive to the nuclear reaction is ‘shifted’ to different depths in
the sample. Thus a depth profile is obtained by scanning a given energy range with the
accelerator and measuring the gamma-yield at each energy. The principle is illustrated in Fig.
2.3 for the case of the 1H(15N,αγ)12C reaction with a well-known resonance at 6.4 MeV. These
measurements are more time consuming than a simple PIGE measurement at one beam
energy.
FIG. 2.3. Principle of depth profiling. Reproduced courtesy of Springer [2.44].
In the following, the basic principles of the analysis of depth profiles are outlined. More
details and special cases are discussed in [2.23].
17
In order to convert the gamma-yield measured at a given energy into a concentration, the
integral of the cross-section over the resonance needs to be calculated. This integral is
essentially equal to the product of the width of the resonance and the cross-section at the
maximum. The energy dependence of a resonance cross-section is given by the Breit-Wigner
formula (in units of area such as barn):
π ⋅ ω ⋅ λ 2 ⋅ Γa Γb
σ ab ( E ) = , (2.8)
Γ2
(E − ER ) + 2
where a and b denote the incoming and outgoing reaction channels, ER is the resonance
energy, is a statistical factor calculated from the spins of the involved nuclear states, Γa and
Γb are the partial widths, Γ is the total width of the resonance and λ is the De-Broglie
wavelength of the incoming particle. A resonance can be characterized, apart from the
resonance energy, by either the maximal cross-section at the resonance energy ER or the total
width Γ or the resonance strength S, which is defined (in units of energy) as:
Γa Γb
S = (2 J + 1) , (2.9)
Γ
where J is the spin of the resonance state. The quantity S is the only way possible to quantify
a resonance in cases where the resonance width is too small for a direct cross-section
measurement. S is, in these cases, determined from an integral measurement over the
resonance, which corresponds to a step function in a yield curve. For details see [2.23].
The method of converting the measured yield into a concentration profile is dependent on the
resonance width, the energy resolution of the beam and the energy straggling of the beam and
involves the integration of these parameters over the energy. This can be done by analysis
codes such as the ERYA code (see Section 5). For cases where the resonance is very narrow,
the beam energy is spread out and the straggling is negligible, the ratio R of the probed atoms
to the total number of atoms in the sample at a beam energy E can be calculated as:
1
R( E ) = Y ( E ) ⋅ ⋅ε , (2.10)
S ⋅η ⋅ ∆Ω
where Y(E) is the gamma-yield, i.e. the number of γ-rays per incident beam ion, measured in
a detector, η and ∆Ω are the efficiency and the solid angle of the detector, respectively, and ε
is the stopping cross-section of the beam ions in the sample at the resonance energy. When the
emitted gamma-radiation is not isotropic but rather has an angular distribution, this must also
be taken into account.
The depth information is obtained from the difference between the resonance energy and the
beam energy at which the gamma-yield is measured, i.e. it is given as an energy loss. The
energy loss can be simply converted to atoms/cm² or to nm (if the density of the sample is
known) by using the stopping power of the beam in the sample; therefore, the sample bulk
composition must be known or measured.
The depth resolution of the method does not depend on the detector resolution, but is
primarily defined by the width of the resonance and can be, in theory, in the nm range or
18
below. In practice, however, other effects such as the beam energy resolution or the Doppler
broadening, which is caused by the thermal motion of the target atoms, can limit the
resolution that is attainable in an experiment.
The energy resolution of low-energy accelerators is typically in the range of 100 eV to 1 keV.
To exploit the full potential of narrow resonance depth profiling, even better beam resolutions
are required [2.26].
The Doppler broadening can be calculated by:
ln 2MEk B T
∆E D = 4 , (2.11)
m
were E is the energy of the beam, kB the Boltzmann constant, T the temperature, M the mass
of the beam ion and m the mass of the sample atom.
In addition, for measurements made deeper within a sample, the energy resolution can
deteriorate due to the energy straggling of the beam. This can be estimated by:
∆E S = 1.2 ⋅ Z 12 Z 2 N , (2.12)
where the straggling ΔE is given in keV, the depth at which a concentration is measured, N is
given in 1018 atoms/cm2, Z1 is the atomic number of the projectile and Z2 is the atomic
number of the sample atoms.
The beam energy distribution, the Doppler broadening and the straggling can all be assumed
to be Gaussian and can therefore be added quadratically. The resonance width has to be added
numerically since the shape of a resonance is a Lorentzian curve. The total experimental
energy resolution ∆Eexp can be converted to depth resolution ∆X by:
∆E exp
∆X = , (2.13)
ε
where ε is the stopping power of the beam in the sample. Resonances suitable for depth
profiling can be found for light nuclei in the mass region up to phosphorus at low beam
energies (below 3 MeV); in fact, almost all light elements have at least one isotope with a low
energy resonance suitable for depth profiling. One prominent example is the 0.43 MeV
resonance of the 15N(p,αγ)12C reaction, which can be used to detect nitrogen in a sample and
which is often used in inverse kinematics to probe hydrogen with an 15N beam. A
comprehensive list of resonances used for depth profiling in material science can be found in
[2.23], where either the cross-sections at maximum and the width or the resonance strength
are tabulated. The energy of the emitted γ-ray is also given in these tables.
Since nuclear reactions can have several resonances, it is to the advantage of the user, when
planning a depth profiling analysis, to find a case where only one resonance contributes to a
depth profile. This is illustrated for the 27Al(p,γ)28Si reaction in Fig. 2.4. Similar data for other
reactions can be found in Section 4 or in the IBANDL data base. For profiling aluminium, the
resonance at 0.992 MeV is the strongest and has a small width. Therefore, it is preferred for
better sensitivity and depth resolution.
19
FIG. 2.4. Yield curve of the 27Al(p,γ)28Si reaction as a function of beam energy [2.27].
The next resonance at higher energies lies at E = 1.002 MeV (seen in Fig. 2.4 as a shoulder of
the 0.992 MeV resonance only), fairly close to the 0.992 MeV resonance, which limits the
energy scan for the 0.992 MeV resonance to a range of 10 keV. This would correspond, for
example, in Si to a depth of around 250 nm.
The next resonance below the 0.992 MeV is at 0.937 MeV. If the layer under investigation is
so thick that a proton beam would lose more than the difference of 55 keV (for silicon that
would correspond to 1.4 µm), then this resonance would influence the depth profile. If the
layer is infinitely thick, all lower resonances lying below 0.992 MeV would contribute to the
depth profiling. The deconvolution of a depth profile that is influenced by more than one
resonance is, in theory, possible, for example by employing the ERYA code, but at the cost of
accuracy. In such cases, it is often better to choose a resonance at low energy. Resonances at
low energies are generally favourable because the background is usually lower and the higher
stopping power leads to a better depth resolution.
2.3. A TYPICAL MINIMAL PIGE FACILITY
In this section, practical instrumental aspects that need to be addressed in the analytical
application of PIGE, but which are also essential, with different constraints, for the
measurement of gamma-ray production cross-sections are presented.
The most common case is for a thick sample to be analysed by proton induced gamma-ray
emission. The mass fraction of the element to be analysed fm (Eq. 2.2) depends on the gamma-
production cross-section and the material stopping power, both normally tabulated as a
function of energy. It also depends on experimental and instrumental quantities, which
include the beam energy (determined by the accelerator beam), the collected charge (also
determined by the accelerator beam), the detector absolute efficiency (determined by the
available detector) and the photon yield of the reaction channel used for detecting the analyte
(also determined by the available detector). Therefore, to perform a PIGE analysis, the
essential set-up would include:
20
(a) A particle accelerator;
(b) A charge collection and measuring system;
(c) A gamma-ray detector (and associated electronics and data acquisition system).
Considering that a PIGE analysis facility may be required to analyze several samples in a row,
maintaining the background level as low as possible throughout these measurements is
essential. Therefore, it is highly recommended to expand the above-mentioned minimal
system by adding two more components:
(i) A beam defining system free of light elements (e.g. Al) that might create background
gamma-radiation (full energy or Compton), associated with a reaction chamber in
vacuum. Such a system, ideal for cross-section measurements, is quite convenient if
small, non-volatile, temperature resistant samples have to be analysed;
(ii) A remotely controlled sample holder.
Items a), b), and c) are elaborated in the following sub-sections.
2.3.1. Particle accelerator
Electrostatic accelerators are the type of accelerators generally adopted in IBA. The reader is
referred to the IAEA Accelerator knowledge portal
(https://nucleus.iaea.org/sites/accelerators/Pages/default.aspx) and the database included
therein for an overview of the characteristics of worldwide electrostatic (mostly) research
accelerators, in regard to the accelerator type/technology (Table 2.2) and the maximum
terminal voltage (Fig. 2.5).
TABLE 2.2. STATISTICS OF ACCELERATOR TYPES AND ACCELERATION TECHNOLOGY

Accelerator Accelerator technology
type Pelletron Tandetron Van de EN-FN- Dynamitron MP EN Other
Graaf MP-UD
Single 3 31 2 8 6
ended
Tandem 47 5 10 28 35 1 2 19
Other 3
*data from IAEA accelerator knowledge portal
Excluding accelerators with low terminal voltage (TV < 200 kV), which are used in industrial
fabrication, and high terminal voltage machines (TV > 6000 kV), which are used mainly in
fundamental nuclear physics, all accelerators with a middle-range terminal voltage (200 kV <
TV < 6000 kV) can be and are used for IBA. For resonance depth profiling in particular, the
range of terminal voltage between 200 and 500 keV is relevant.
To a large extent, they also fulfill the requirements of a PIGE analysis. A reliable energy
control system, e.g. the Generating VoltMeter (see Section 3.1) and a lower limit for the
energy of protons around 200–500 keV are needed if the experimental set-up is dedicated to
the measurement of PIGE cross-sections and high resolution PIGE depth profiling (see
Section 2.2). Beam currents up to a few hundred nA can suit the large majority of specific
21
PIGE requirements and are easily achieved with standard ion sources, e.g. duoplasmatron and
sputtering ion sources.
FIG. 2.5. The maximum terminal voltage featured by the accelerators listed in the IAEA accelerator
knowledge portal (data on January 2016).
2.3.2. Charge collection and measuring system
In PIGE analysis, the proton energy will be almost always below 5 MeV and the current will
be limited to a few hundred nA for a power around 1 W to be dissipated. Such beam currents
can be easily measured, in vacuum, by simple insertion devices, such as (i) a transmission
Faraday cup (TFC), which intercepts the beam partially and periodically, before or after the
sample, and (ii) a Faraday cup (FC) that intercepts the beam continuously after the sample
[2.28]. The current can be integrated from both devices by a current digitizer (there are many
commercial models) that can give a pulse with a lower limit of 0.1 nC, which guarantees, in
theory, a precise measure of the collected charge. Besides the problem connected to the
emission of secondary electrons from the target (see Section 3.4), there are problems
connected with the device itself. The first is that the FC, in PIGE set-ups and, in general IBA
set-ups, is located close to the sample, which often forms a unit with the sample holder, and is
generally not shielded. This may create a gamma-ray background that affects the detector
response. Light materials, like carbon, must be avoided and preference given to high Z metals.
The second and more important problem is produced by the emission of secondary electrons
from the FC after the proton impact. The secondary electron energy is a few ten eV and the
flux is proportional to the cosine of the emission angle with respect to the surface normal.
Their escape increases the positive current in the circuit and generates a systematic
overestimation of the charge. Secondary emission is reduced by assuring that the cup shape
features a diameter as small as possible compared to the length of the FC and is efficiently
suppressed by introducing a ring in front of the FC, which is maintained at a negative
22
potential of a few hundred volts, to push the electron back to the FC surface. The same
suppressor can be put in action in the case of a beam profile monitor (BPM).
In external beams a relative charge monitor can be envisaged. It is a commonly-used device

that is preferred over a beam chopper (see Section 3.4) because of its mechanical simplicity,
which takes advantage of the in-beam stability of the Si3N4 membranes that are nowadays
widely used as exit windows. For each energy, the number of X-rays emitted by Si in the
membrane, and collected by a fixed Si(Li) or Silicon Drift Detector (SDD) dedicated detector,
is in constant proportion with the charge that has been transmitted through it and delivered to
the sample. In the case of a standard-based method, the relative charge monitor, given by the
X-ray counts, assures the comparison with the standard. For a standardless analysis, the
system must be carefully calibrated, energy by energy, against known collected charges,
which is not easy for a system in air.
2.3.3. Gamma-ray detector
Both the error (statistical + fit) on the gamma-yield, as well as the PIGE analysis minimum
detection limit (MDL) are lower when the yield is higher and the background (from peaks
overlap or competing reactions, or from Compton interactions of higher energy radiation with
the detector) is lower in the energy region (normally the Full Width at Half Maximum)
covered by the gamma-ray peak used to detect the analyte. This puts some minimum
requirements on the efficiency and energy resolution of gamma-ray detectors and on the
experimental set-up.
Taking into consideration the elements from lithium to phosphorus and the reactions listed in
Table 2.1, it is expected that characteristic gamma-rays below or around 2 MeV will be
detected, with a difference between nearest neighbouring lines from 2 to 343 keV.
Background peaks come from the 232Th series (208Tl, 212Pb, 212Bi, 224Ra, 228Ac), the 238U series
(214Pb, 214Bi, 226Ra) and 40 K. The most intense peaks that are expected inside an accelerator
vault are from 208Tl (510.7 and 583.2 keV), 212Pb (238.6 keV), 214Pb (295.2 keV), 214Bi (609.3
and 1120.3 keV), 228Ac (911.2 and 968.9 keV) and 40 K (1460.8 keV). High volume Ge
detectors can provide a resolution at FWHM of 1.2–1.5 keV at Eγ = 59 keV and of 2–3 keV at
Eγ = 1332 keV, which is sufficient to separate all the expected PIGE peaks below 1636 keV
from each other and from the background peaks, with the exception of the 585 keV line from
25
Mg and the 1459 keV line from 19F.
Modern, fast scintillator detectors, like the LaBr3, give a FWHM energy resolution roughly
between 7% at 60 keV and 2% at 1460 keV, which means an absolute value between 8.4 keV
at 60keV and 58 keV at 1460 keV. The performance is acceptable for energies below about
800 keV, where the self-activity of LaBr3 is not interfering. Self-activity comes from the
naturally occurring 138La that decays to 138Ba by emitting a gamma-ray of 1436 keV (66.4%)
or to 138Ce by emitting a gamma-ray of 789 keV (66.4%). Since they are compact and light,
the handling of these detectors is quite easy in comparison to a Ge detector with its liquid
nitrogen dewar.
The yield of a gamma-ray at energy Eγ depends on the absolute efficiency of the detector εabs
[2.29]. Assuming a point-like gamma-ray source, the detector absolute efficiency εabs is
defined as the ratio between the gamma detected, at their full energy Eγ, and the gamma
emitted in the whole solid angle in the same time, i.e.:
23
N Det (T ) 1
N TOT (T ) 4π ∫∫
ε abs = = ∆θ∆ϕε ( E γ , θ , ϕ ) sin θdθdϕ , (2.14)
where ε(Eγ, θ, φ) is the detector efficiency in its more general form, i.e. as a function of the
photon flight direction (θ, φ) and energy Eγ. Absolute efficiency increases when the solid
angle covered by the detector is wider but also when the detector entrance window is thinner.
The absolute efficiency is measured (see methodological details in Section 3.2) with the use
of calibrated sources. The most often-used combination includes 133Ba, 152Eu, 137Cs, 60Co,
which results in a relevant number of sufficiently intense, well-spaced peaks, from 53.1622
keV (133Ba) to 1408.013 keV (152Eu), that covers both the low energy region, where the effect
of the entrance window may strongly reduce εabs, and the high one.
The absolute efficiency of a 5cm × 5cm LaBr3 detector is not much lower than that of a 6cm ×
6cm Ge (see Fig. 2.6). However the MDL, usually defined as the analyte quantity that gives a
number of counts corresponding to , where NBKG is the number of background counts
below the gamma-ray peak, is reduced in a Ge detector, given its superior energy resolution.
This makes Ge detectors an excellent choice for cross-section measurements or application of
PIGE for materials characterization.
FIG. 2.6. The absolute efficiency, as a function of gamma energy, of: a) a 6cm × 6cm REGe detector
and b) a 5cm × 5cm LaBr3 scintillator at the same distance from the gamma-ray source.
The achievement of best detector performance requires some attention. For increasing the
integral yield/charge ratio, which helps lowering MDL’s or shortening measure time, the solid
angle may be increased by moving the detector towards the target. When deciding on the final
distance from the target, a compromise must be reached between solid angle, increased count
rate, pile-up distortion and possible increase in background due to the electromagnetic
cascade, which originates from gamma-rays generated outside the target. A different solution
consists in using a multi-detector system that will detect photons under the sum of the solid
angles for single detectors. Although it is an expensive solution, if the investment has already
24
been made, e.g. to measure in shorter times angular distributions, and the gamma-emission
from the examined reaction is expected to be quite isotropic, the effort spent in combining the
individual spectra may be compensated by the increased overall detection efficiency.
2.3.4. Example of a minimal PIGE set-up
A minimal PIGE set-up is shown in Fig. 2.7. A general purpose external micro-beam line,
already provided with a PIXE and RBS detector system, as well as with a relative charge
monitor that is based on the X-ray emission from the Si3N4 exit window, has been
implemented with a LaBr3 scintillator for the analysis, in air, of samples that cannot stand
vacuum conditions, e.g. large objects (see Section 2.4.1). The addition of a detector for PIGE,
that does not require cooling and is small and light, simplifies the operation and produces an
efficient set-up for measuring characteristic gamma-rays that are below about 800 keV and
measurements that can be performed with a standard-based method or, as a second choice,
can be semi-quantitative, given the difficulty of a monitor calibration.
FIG. 2.7. The external microbeam set-up at CMAM-UAM of Madrid, Spain implemented with a LaBr3
detector.
2.3.5. Example of an in-vacuum multi detector system
PIGE analysis and, even more so, the measurement of cross-sections may require a more
elaborated experimental set-up, e.g. the one shown in Fig. 2.8. In this case, the PIGE reaction
is produced inside a vacuum chamber, equipped with three high purity Ge detectors for
gamma-rays as well as a large surface Si(Li) detector and a fast SDD detector for X-rays.
Such a system has the capability of measuring differential PIGE cross-sections at three angles
simultaneously, allowing the determination of angular distribution asymmetries and the
production of more precise total cross-sections. On the analytic side, the fact that this set-up,
like the previous in-air one, features both gamma-ray and X-ray detectors is of great
advantage for fully determining the target composition, over a large Z range, by combining
PIGE and PIXE data. It can only be used for samples, however, that can stand the vacuum
without degrading.
25
FIG. 2.8. The PIGE set-up at LABEC-INFN in Florence, Italy with the three germanium detectors in
the forefront.
In summary, implementing a PIGE experimental station is relatively simple and can be done
at a reasonable cost. A high resolution detector, such as an ultrapure germanium detector,
will be able to solve the large majority of analytic problems, given that it has the efficiency
and energy resolution necessary to univocally identify the light elements against other
elements in the sample or the background signals. The best use of PIGE will certainly be in
combination with PIXE, due to the fact that these two techniques are commensurate for the
analysis of elements such as Na, Mg, Al and Si and that PIGE can be further extended to
low Z in order to cover the lighter elements.
2.4. EXAMPLES OF APPLICATIONS IN NA BULK ANALYSIS
2.4.1. Identification of lapis-lazuli in paint layers
Lapis-lazuli is a blue pigment that has been used since the Middle Ages in different kinds of
painting techniques. It is obtained from a semi-precious stone that is mainly composed of
lazurite (3Na2O⋅3Al2O3⋅6SiO2⋅2Na2S), but that also contains small amounts of calcite
(CaCO3) and pyrite (FeS2) in variable quantities. It was considered to be a very precious and
expensive pigment; therefore, it was used, in particular, for religious-themed art, where it
emphasized the symbolic value of the work. In canvas and wood paintings, the pigment can
be regularly found mixed with lead white (2PbCO3⋅Pb(OH)2) to produce a paler shade.
Although PIXE is a useful analytical technique to distinguish different pigments [2.30, 2.31],
the effectiveness of the technique is severely limited by different factors in cases where lapis-
lazuli is mixed with lead white. First, the presence of large quantities of lead results in a
strong self-absorption of the low energy X-rays emitted by the light elements (Na, Al and Si)
characteristic of this pigment within the paint layer itself. In addition, Al and Si cannot be
considered a unique fingerprint for lapis-lazuli, since they are found in many other pigments
that are often added to the paint layer to modify the chromatic shade. Furthermore, the
presence of the protective layer of varnish, which is made from organic materials and placed
with varying thickness on wood and canvas paintings, further reduces the detection sensitivity
for these light elements. Finally, the highest Z element of lazurite (S), which is less affected
by absorption problems, is difficult to identify in the presence of a large amount of lead, due
26
to the overlap of Pb M lines with the S Kα line. A deconvolution of the K line of S from the
M lines of Pb is practically impossible, even with the good energy resolution of SDD
detectors.
To solve this problem, PIGE was used to identify lapis-lazuli in pigment layers by detecting
the 441 keV gamma-ray from the 23Na(p,p'γ)23Na reaction [2.32]. This is shown in the
following example from the study of the ‘‘Madonna dei fusi’’ (ex-Redford version, 1501,
private collection), which is a famous painting by Leonardo da Vinci (Fig. 2.9).
Measurements were performed at the KN3000 Van de Graaff accelerator of INFN in

Florence, using a proton beam extracted into atmosphere, which had 2.7 MeV of energy on
the target, with a 1 mm diameter size and intensity ranging from some tens of pA to 0.1 nA, in
order to avoid any risk of damage. Two Si(Li) detectors were used to detect X-rays for PIXE,
while a planar Ge detector was used for PIGE (23 mm thickness, 700 eV FWHM energy
resolution at 122 keV, placed at about 135° to the beam direction and covering a solid angle
of about 0.5 sr).
FIG. 2.9. On the left, the ‘‘Madonna dei fusi’’ painting from Leonardo da Vinci during PIXE–PIGE
measurements. On the right, PIGE spectra obtained irradiating two blue areas in: (a) a restored area;
(b) an original area. The presence of lapis-lazuli pigment in the original paint layer is evidenced by
the 441 keV γ-ray peak from the 23Na(p,p′γ)23Na reaction [2.33].
PIXE spectra obtained from some of the blue areas revealed the presence of zinc white (ZnO)
and cobalt blue (CoO⋅Al2O3), which only came into use at the beginning of the 19th century.
It was thus possible, using these PIXE measurements, to identify many areas that had
undergone restoration (mainly in the mantle of the Virgin).
In the original blue areas, which are characterized by a large amount of Pb and were clearly
associated with lead white, none of the elements characterizing a blue pigment were detected
by PIXE; not even Cu from azurite (copper carbonate, Cu3(CO3)2(OH)2), the most probable
period-compatible alternative blue pigment. Therefore, if conclusions were based on PIXE
measurements alone, the use of lapis-lazuli for these blue zones could only be indirectly
27
suggested. PIGE measurements, however, positively revealed the presence of lapis-lazuli
through the detection of the 441 keV gamma-rays from the 23Na(p,p′γ)23Na reaction shown in
Fig. 2.9b.
2.4.2. Quantification of Na in ancient glasses
It is well known that a typical effect of ancient glass ageing [2.34] is surface alteration due to
a water-inflicted corrosion processes. A surface decrease (leaching) in alkali ion
concentration, due to exchange between H+ from water and Na+ or K+ in glass, is caused by
humidity from the environment. Therefore, a layer of hydrate silicates is formed in the altered
surface, which is depleted in Na, K or Ca elements and enriched in hydrogen. The extent of
the weathering process depends on time and on burial/exposure conditions, with leached
layers extending up to some tens of µm [2.35, 2.36].
Detection and quantification of Na in ancient glasses can be crucial, for instance to distinguish
the natron glass of the Roman times (with high Na2O and low K2O and MgO content) from
later ones like plant ash glass, which have been used since the Middle Ages (with low Na2O
and high K2O content) [2.37, 2.38]. However, the presence of Na in the bulk may be
completely hidden to a PIXE analysis, since the altered layers present at the surface may
strongly absorb the X-rays from the lightest elements and thus heavily affect the
characterization of the bulk material.
To address problems of this kind, a combined use of PIXE and PIGE is particularly effective
[2.25, 2.39–2.43], since it exploits the 23Na(p,p′γ)23Na reaction that produces the 441 keV
gamma-ray for Na determination. PIXE information on Na is limited to the first few microns
from the surface due to strong self-absorption effects (Na X-ray energy is 1.04 keV), while
the integrated yield of gamma-rays refers to a much larger depth, since gamma-rays are
detected with no attenuation regardless of the depth from where they originate. It is thus
possible to immediately qualitatively distinguish between surface and bulk composition.
An example from the PIXE-PIGE measurement of an area of a blue coloured tessera from the
famous Roman complex of Villa Adriana (Tivoli) [2.40] is shown in Fig. 2.10. The examined
area displays three different kinds of surfaces with different characterizations: a brighter
colour (region 1), a more opaque area (region 2) and a white, clearly altered area (not shown
in Fig. 2.10). The measurements were carried out at the 3 MV Tandetron of INFN in Florence
at the external microprobe set-up, using a proton beam of 2.9 MeV energy on target extracted
into atmosphere through a 100 nm Si3N4 window, with 100 µm size and about 1 nA intensity.
The beam was scanned over an area of 2 × 2 mm2 and larger areas were analysed by
combining beam scans on adjacent zones obtained by mechanical target displacement. Two
detectors were employed for PIXE, a SDD and a Si(Li), which provided altogether a good
efficiency from Na X-rays to those of the highest-Z elements, while a HPGe detector was
used for PIGE (n-type, 25% efficiency, placed 6 cm from target at 0°).
28
FIG. 2.10. X-ray (left) and γ-ray spectra (right) referring to two of the three different regions of a blue
glass tesserae. Only the energy ranges around the Na peaks are displayed [2.33].
The differences between PIXE and PIGE spectra observed in Fig. 2.10 can be explained by
the presence of surface alterations leading to Na-depleted layers. Both X-rays and gamma-
rays are clearly detected in region 1. In region 2, Na X-rays are not detected; however, there is
no appreciable difference in the gamma-ray yields of the two regions. This indicates that the
opaque area has a Na-depleted layer that is very thin. Since Na X-rays are not totally
undetected, the depletion layer in region 2 cannot be larger than a few microns. Therefore, the
fraction of missing gamma-ray yield due to proton energy loss in the thin Na-depleted surface
layer is not relevant. A quantitative evaluation of Na content from PIGE spectra from region 2
alone would therefore provide a reasonable lower limit, but quantification through PIXE
would be largely incorrect.
In the specific example, the Na2O content derived from PIGE analysis using the standard-
based method, as described in Section 2.1 above, was around 15%. This allows the
unambiguous identification of the Roman origin of the studied glass tesserae, since, in the
literature, there is evidence that the composition of natron glasses during Roman times was
almost constant, with a Na2O concentration between 16% and 18%.
2.5. HYDROGEN DEPTH PROFILING
Hydrogen depth profiling by using a narrow resonance in a particle-induced gamma-ray

emission reaction is of importance because it offers a straightforward, easy-to-quantify and
depth-sensitive method for the investigation of hydrogen in materials, which cannot be
achieved with most other analytical techniques. The difference between this method and the
other PIGE measurement methods is that the beam that is employed is not a proton beam but
rather a heavier beam such as 15N, and the target is the lighter hydrogen atom 1H. Therefore,
the proton-induced gamma-ray emitting reaction is used in inverse kinematics. The method
works also with 18O or 19F ions as projectile beams, but the resonance in 1H(15N,αγ)12C at 6.4
29
MeV has favourable features regarding sensitivity and depth resolution (see Section 2.2 for
details) and is most frequently used for hydrogen depth profiling. With this reaction, depth
resolutions of a few nm and sensitivities down to the ppm range have been achieved [2.44].
Recent reviews of the method can be found in [2.45] and [2.46].
The principle of this method has already been described in Section 2.2 above. However,
having the heavy beam particle is an advantage since the background from other components
of samples is negligible and the emitted gamma-ray of 4.4 MeV from the reaction is well
separated from room background. Therefore, an optimum resolution of the gamma-detector is
not needed, i.e. NaI(Tl) or BGO detectors can be used, which offer a high gamma-ray
efficiency and, consequently, a higher sensitivity. Because of the high gamma-ray efficiency,
smaller beam currents can be used, which is preferred given that beam-induced changes in the
sample and in the hydrogen depth profile can be quite large when using a 15N beam compared
to a proton beam.
There are several laboratories equipped with a dedicated set-up for hydrogen depth profiling.
They are either optimized with regard to sample handling, maintenance and costs enabling a
high rate of processing samples, or high sensitivity. Figure 2.11 shows a set-up dedicated to
the detection of low hydrogen concentrations, as used at the University of Bochum (RUB).
The sample holder assembly is mounted on the end of a pipe, which can be surrounded by a
12 × 12 inch NaI(Tl) detector in 4π geometry, to gain a detection efficiency for the 4.4 MeV
gamma-ray of almost 50 %.
Figure 12 gives an example of a depth profile for a multilayer sample, as used in research for
future hydrogen storage materials that are loaded with a high hydrogen concentration. The
layer structure has been determined by RBS and is sketched at the bottom. There is a surface
peak of hydrogen, which is observed for almost all samples handled within the laboratory
environment. In the region of the titanium layer that is partially oxidized, the hydrogen
concentration increases as the oxygen content decreases.
30
FIG. 2.11. Set-up for hydrogen depth profiling. Reproduced courtesy of Springer [2.45].
31
70
60
50
H concentration [atom %] 40
30
20
10
Pd Ti+O Ti Cr Si3N 4 SiO2

0
-10
-200 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200
depth (1e15 at/cm²]
FIG. 2.12. Hydrogen depth profile of a multilayer system. The layer structure is indicated at the
bottom.
Hydrogen can alter the properties of materials drastically, even at low concentrations. An
example is hydrogen in olivine, which is the most abundant mineral in the earth’s crust. In
this instance, hydrogen concentrations in the ppm range have a strong influence on
mechanical properties such as the plastic flow in geological conditions. For such applications,
there have been significant efforts to improve the sensitivity of the method into the ppm
range.
When hydrogen concentrations below 100 ppm have to be measured, special conditions have
to be fulfilled. One is that the constructive elements of the set-up, such as collimators, which
are hit by the beam, have to have low hydrogen contaminations and have to be mounted as far
from the detector as possible. Another condition is that the surface of the sample has to be
free of hydrogen, which is usually easily adsorbed onto almost any sample surface. When a
low hydrogen concentration in the depth of a material is measured, the beam has to hit the
sample at an energy above the resonance energy (see Section 2.2). The cross-section for the
reaction at the surface is several (almost five) orders of magnitude lower than at the peak of
the resonance. However, if the hydrogen concentration of a surface layer is five orders of
magnitude higher than in the bulk, the number of γ-rays from the surface is equal to
contributions of the bulk. It should be noted that this problem arises even if the surface layer
is only a few atomic layers thick.
Therefore, a UHV sample chamber that can produce a vacuum better than 10-8 mbar and
options to clean sample surfaces in-situ are mandatory for a facility intending to do low
concentration hydrogen depth profiling analysis. In the set-up of Fig. 11, for example, surface
cleaning can be accomplished by heating the sample via electron bombardment with an ion
sputter gun. Details can be found in [2.47].
32
[2.1] SIPPEL, R.F., GLOVER, E.D., Sedimentary rock analysis by charged particle
bombardment, Nucl. Instrum. Methods 9 (1960) 37–48.
[2.2] PIERCE, T.B., PECK, P.F., HENRY, W.M., Determination of carbon in steels by
measurement of the prompt γ-radiation emitted during proton bombardment, Nature
204 (1964) 571.
[2.3] PIERCE, T.B., PECK, P.F., HENRY, W.M., The rapid determination of carbon in
steels by measurement of the prompt radiation emitted during deuteron
bombardment, Analyst 90 (1965) 339.
[2.4] PIERCE, T.B., PECK, P.F., CUFF, D.R.A., Application of inelastic proton scattering
to the rapid determination of silicon in steels, Anal. Chim. Acta 39 (1967) 433.
[2.5] BORDERIE, B., BARRANDON, J.N., New analytical developments in prompt
gamma- ray spectrometry with low-energy tritons and alpha particles, Nucl. Instrum.
Methods 156 (1978) 438.
[2.6] BORDERIE, B., Present possibilities for bulk analysis in prompt gamma-ray
spectrometry with charged projectiles, Nucl. Instrum. Methods 175 (1980) 465.
[2.7] BONI, C., CEREDA, E., BRAGA MARCAZZAN, G.M., DE TOMASI, V., Prompt
gamma emission excitation functions for PIGE analysis, Nucl. Instrum. Methods
Phys. Res. B 35 (1988) 80.
[2.8] BONI, C., CARUSO, E., CEREDA, E., BRAGA MARCAZZAN, G.M., and
REDAELLI, P., A PIXE-PIGE system for the quantitative elemental analysis of thin
samples, Nucl. Instrum. Methods Phys. Res. B 40/41 (1989) 620.
[2.9] BONI, C., CARUSO, E., CEREDA, E., BRAGA MARCAZZAN, G.M.,
REDAELLI, P., Simultaneous PIXE-PIGE analysis of thin and thick samples, Nucl.
Instrum. Methods Phys. Res. B 49 (1990) 106.
[2.10] ANTTILA, A., HÄNNINEN, R., RÄISÄNEN, J., Proton-induced thick-target
gamma-ray yields for the elemental analysis of the Z = 3–9, 11–21 elements, J.
Radioanalytical Chemistry 62 (1981) 293.
[2.11] RÄISÄNEN, J., HÄNNINEN, R., Heavy element (Z > 30) thick-target gamma- ray
yields induced by 1.7 and 2.4 MeV protons, Nucl. Instrum. Methods 205 (1983) 259.
[2.12] LAPPALAINEN, R., ANTTILA, A., RÄISÄNEN, J., Absolute α-induced thick-
target gamma-ray yields for the elemental analysis of light elements, Nucl. Instrum.
Methods 212 (1983) 441.
[2.13] KISS, Á.Z., KOLTAY, E., NYAKÓ, B., SOMORJAI, E., ANTTILA, A.,
RÄISÄNEN, J., Measurements of relative thick target yields for PIGE analysis on
light elements in the proton energy interval 2.4–4.2 MeV, J. Radioanalytical Nucl.
Chemistry 89 (1985) 123.
[2.14] RÄISÄNEN, J., WITTING, T., KEINONEN, J., Absolute thick-target γ-ray yields
for elemental analysis by 7 and 9 MeV protons, Nucl. Instrum. Methods Phys. Res.
B 28 (1987) 199.
[2.15] KISS, A. Z., I. BIRON, T. CALLIGARO, J SALOMON, Thick target yields of
deuteron induced gamma-ray emission from light elements, Nucl. Instrum. Methods
Phys. Res. B 85 (1994) 118.
[2.16] ELEKES, Z., KISS, A.Z., BIRON, I., CALLIGARO, T., SALOMON, J., Thick
target γ-ray yields for light elements measured in the deuteron energy interval of 0.7–
3.4 MeV, Nucl. Instrum. Methods Phys. Res. B 168 (2001) 305
[2.17] MATEUS, R., JESUS, A.P., BRAIZINHA, B., CRUZ, J., PINTO, J.V., RIBEIRO,
J.P., Proton-induced γ-ray analysis of lithium in thick samples, Nucl. Instrum.
Methods Phys. Res. B 190 (2002) 117
33
[2.18] MATEUS, R., JESUS, A.P., RIBEIRO, J.P., Quantitative analysis of light elements
in thick samples by PIGE, Nucl. Instrum. Methods Phys. Res. B 219–220 (2004)
519.
[2.19] MATEUS, R., JESUS, A.P., RIBEIRO, J.P., A code for quantitative analysis of light
elements in thick samples by PIGE, Nucl. Instrum. Methods Phys. Res. B 229 (2005)
302.
[2.20] CACIOLLI, A., CHIARI, M., CLIMENT-FONT, A., FERNÁNDEZ-JIMÉNEZ, M.
T., GARCÍA-LÓPEZ, G., et al., Measurements of γ-ray emission induced by protons
on fluorine and lithium, Nucl. Instrum. Methods Phys. Res. B 249 (2006) 98.
[2.21] CACIOLLI, A., CALZOLAI, G., CHIARI, M., CLIMENT-FONT, A., GARCIA, G.,
et al., Proton elastic scattering and proton induced γ-ray emission cross-sections on
Na from 2 to 5 MeV, Nucl. Instrum. Methods Phys. Res. B 266 (2008) 1392.
[2.22] CALZOLAI, G., CHIARI, M., LUCARELLI, F., NAVA, S., PORTARENA, S.,
Proton induced γ-ray emission yields for the analysis of light elements in aerosol
samples in an external beam set-up, Nucl. Instrum. Methods Phys. Res. 268 (2010)
1540.
[2.23] WANG, Y, NASTASI, M. (Eds), Handbook of Modern Ion Beam Materials
Analysis, 2nd edn, Materials Research Society, Pennsylvania (2009).
[2.24] MATEUS, R. A., JESUS, P., FONSECA, M., LUÍS, H., RIBEIRO, J.P., The
sensitivity of the PIGE analytical technique, Nucl. Instrum. Methods Phys. Res.
B 264 (2007) 340.
[2.25] CLIMENT-FONT, A., MUNOZ-MARTIN, A., YNSA, M.D., ZUCCHIATTI, A.,
quantification of sodium in ancient Roman glasses with ion beam analysis, Nucl.
[2.26] SCHULTE, W.H., EBBING, H., BECKER, H.W., BERHEIDE, M., BUSCHMANN,
M., et al., High resolution depth profiling in near-surface regions of solids by narrow
nuclear reaction resonances below 0.5 MeV with low energy spread proton beams,
Vacuum 44 (1993) 185.
[2.27] HARISSOPULOS, S., CHRONIDOU, C., SPYROU, K., PARADELLIS, T.,
ROLFS, C., et al., The 27Al (p,gamma)28Si reaction: direct capture cross-section and
resonance strengths at Ep = 0.2–1.12 MeV, Eur. Phys. J. A 9 (2000) 479.
[2.28] A. WU CHAO, A., TIGNER, M. (Eds), Handbook of accelerator Physics and
Engineering, World Scientific (2002), pp. 556–559.
[2.29] ZUCCHIATTI, A., CORVISIERO, P., PIGE: a useful technique for the study of
innovative energy related materials. Technical note on the measurement of gamma-
ray production cross-sections, Energy Procedia 41 (2013) 110–118.
[2.30] MANDÒ, P.A., FEDI, M.E., GRASSI, N., MIGLIORI, A., Differential PIXE for
investigating the layer structure of paintings, Nucl. Instrum. Methods Phys. Res.
B 239 (2005) 71.
[2.31] GRASSI, N., BONANNI, P., MAZZOTTA, C., MIGLIORI, A., MANDÓ, P.A.,
PIXE analysis of a painting by Giorgio Vasari, X-Ray Spectrom. 38 (2009) 301.
[2.32] GRASSI, N., MIGLIORI, A., MANDO, P.A., CALVO DEL CASTILLO, H.,
Identification of lapis-lazuli pigments in paint layers by PIGE measurements, Nucl.
Instrum. Methods Phys. Res. B 219–220 (2004) 48.
[2.33] CHIARI, M., private communication.
[2.34] DAL BIANCO, B., BERTONCELLO, R., MILANESE, L., BARISON, S., Glasses
on the seabed: surface study of chemical corrosion in sunken Roman glasses, J. Non-
Cryst. Solids 343 (2004) 91.
[2.35] WEBER, G., STRIVAY, D., MARTINOT, L., GARNIR, H.P., Use of PIXE– PIGE
under variable incident angle for ancient glass corrosion measurements, Original
Nucl. Instrum. Methods Phys. Res. B 189 (2002) 350.
34
[2.36] GRATUZE, B., BARRANDON, J.N., DULIN, L., AL ISA, K., Ancient glassy
materials analyses: a new bulk nondestructive method based on fast neutron
activation analysis with a cyclotron, Nucl. Instrum. Methods Phys. Res. B 71 (1992)
70.
[2.37] SAYRE, E.V., SMITH, R.W., Compositional categories of ancient glass, Science
133 (1961) 1824.
[2.38] HENDERSON, J., The Science and Archaeology of Materials, Routledge, London
and New York (2000).
[2.39] GÓMEZ-TUBÍO, B., ONTALBA SALAMANCA, M.A., ORTEGA-FELIU, I.,
RESPALDIZA, M.A., AMORES CARREDANO, F., GONZÁLEZ-ACUÑA, D.,
PIXE–PIGE analysis of late Roman glass fragments, Nucl. Instrum. Methods Phys.
Res. B 249 (2006) 616.
[2.40] GRASSI, N., GIUNTINI, L., MANDÒ, P.A., MASSI, M., Advantages of scanning-
mode ion beam analysis for the study of Cultural Heritage, Nucl. Instrum. Methods
Phys. Res. B 256 (2007) 712.
[2.41] ŠMIT, Z., JEZERŠEK, D., KNIFIC, T., ISTENIC, J., PIXE–PIGE analysis of
Carolingian period glass from Slovenia, Nucl. Instrum. Methods Phys. Res. B 267
(2009) 121.
[2.42] ŠMIT, Z., TARTARI, F., STAMATI, F., VEVECKA PRIFTAJ, A., ISTENIC, J.,
Analysis of Roman glass from Albania by PIXE–PIGE method, Nucl. Instrum.
Methods Phys. Res. B 96 (2013) 7.
[2.43] ŠMIT, Z., MILAVEC, T., FAJFAR, H., REHREN, LANKTON, J.W., GRATUZE,
B., Analysis of glass from the post-Roman settlement Tonovcov grad (Slovenia) by
PIXE–PIGE and LA-ICP-MS, Nucl. Instrum. Methods Phys. Res. B 311 (2013) 53.
[2.44] ENDISCH, D., STURM, H., RAUCH, F., Nuclear reaction analysis of hydrogen at
levels below 10 at.ppm, Nucl. Instrum. Methods Phys. Res. B 84 (1994) 380.
[2.45] FRITZSCHE, H., HUOT, J., FRUCHART, D. (Eds), Neutron Scattering and Other
Nuclear Techniques for Hydrogen in Materials, Springer (2016).
[2.46] WILDE, M., FUKUTANI, K., Hydrogen detection near surfaces and shallow
interfaces with resonant nuclear reaction analysis, Surf. Sci. Rep. 69 (2014) 196.
[2.47] TRAEGER, F., KAUER, M., WÖLL, CH., ROGALLA, D., BECKER, H.W.,
Analysis of surface, subsurface, and bulk hydrogen in ZnO using nuclear reaction
analysis, Phys. Rev. B 84 (2011) 075462.
35
3. METHODOLOGY OF CROSS-SECTION MEASUREMENTS
Accurate differential cross-section determination is dependent on the accuracy of recordings

and measurements performed in the laboratory. These measurements depend on a number of
parameters, which will be described in Sections 3.1–3.4, including (i) energy calibration of
the accelerator, (ii) efficiency calibration of the detector, (iii) target preparation and (iv)
charge collection. Since these parameters all directly affect the cross-section calculations, it is
of paramount importance that they are all undertaken with great care for accuracy and
precision so as to minimize error and discrepancy. Further details on the methodology in
determining cross-sections are discussed in Section 3.5.
3.1. ENERGY CALIBRATION OF THE ACCELERATOR
Since the correct recording of the precise energy at which the reaction has occurred is vital to
the accurate determination of differential cross-sections, especially when narrow resonances
are involved, an energy calibration of the accelerator is performed before any measurements
are made. It is also advisable to confirm results by performing a second energy calibration
after all of the measurements have been made. There are several possible calibration methods
proposed in the literature, three of which are described below in more detail. There are also
several ways of measuring the energy of the accelerator during the experiment [3.1]. The most
common method is by measuring the voltage in the terminal of the accelerator using a
Generating VoltMeter (GVM). Another possibility is by measuring the magnetic field of a
bending magnet with the aid of an NMR probe.
3.1.1. Narrow resonances
One of the most accurate and widely used methods of calibration employs gamma-ray
reactions that exhibit well-studied, strong, narrow resonances [3.1, 3.2]. Most of these
resonances are produced when a proton incident beam is used. A list of commonly used
resonances can be found in Table 3.1. The beam is directed onto a target containing the
appropriate element. It is considered good practice to choose a thick, polished, mono-
elementary target, but a composite one could be used as long as the composition was known.
The emitted gamma-rays are recorded by a detector, such as a HPGe or NaI detector, placed
at a fixed angle with respect to the beam axis.
TABLE 3.1. COMMON RESONANCES USED FOR ENERGY CALIBRATION

Reaction Resonance energy Width Gamma-ray emitted Reference
(keV) (keV) (keV)
27
Al(p,γ)28Si 991.9 0.1 1779 [3.3]
13
C(p,γ)14N 1746.9 0.075 2730 [3.4]
32
S(p,p′γ) 32S 3379 1.2 2230 [3.5]
14
N(p,p′γ)14N 5937 17 2313 [3.2]
The calibration procedure is as follows: (i) the accelerator terminal is set at a voltage, or the
magnet is set at an NMR value, corresponding to an energy lying well below the resonance.
After accumulating a certain amount of integrated beam charge, (ii) the gamma-rays emitted
by the reaction are recorded. Then (iii) the voltage of the terminal is increased and the
procedure is repeated, making sure to keep the same beam charge. As the beam energy
37
approaches the resonance energy, (iv) the number of gamma-rays that are emitted increases as
the tail of the resonance enters the surface of the target. Further increment of the energy leads
to further increment of the production of the gamma-rays as a larger part of the resonance
contributes to it, until (v) a plateau is reached. This plateau indicates that the resonance has
fully entered the target and no further increment of gamma-rays is expected. (vi) The
experimental data is fitted using a sigmoidal function and the NMR value that corresponds to
the resonance energy is given by sigma. An example of the resulting graph (yield vs terminal
voltage (NMR value)) is presented in Fig. 3.1.
FIG. 3.1. Yield of gamma-rays as a function of terminal voltage (NMR value) as performed for the
27
Al(p,γ)28Si resonance from Table 3.1 above. Experimental data points are fitted with a sigmoidal
function.
(vii) Steps i–vi are repeated with another resonance of another isotope. (viii) Finally, the
terminal voltages (NMR values) are plotted against the resonance energies and the data is
fitted with a linear equation. This linear equation represents the accelerator’s calibration.
3.1.2. Neutron producing threshold reactions
A second method of calibrating the energy of the accelerator is based on neutron producing
threshold reactions [3.6–3.8]. The underlying principle is that when the energy of the beam is
lower than the threshold, no reaction occurs and no neutrons are produced. Once again, a
thick target containing the neutron producing element is needed. A neutron counter, such as
BF3, is used for the detection of neutrons. The terminal voltage of the accelerator is set to a
lower starting value than the threshold for the reaction and the produced neutrons are counted.
Since the energy is lower than the threshold, any neutrons that exist originate mainly from
other sources such as reactions of other elements, the building materials of the chamber or the
38
collimators. Since the cross-sections of secondary reactions increase with voltage, it can be
deduced that the number of emitted neutrons increases linearly with the increase of voltage.
When the threshold energy is reached, the production of neutrons increases rapidly, but the
relationship remains linear. On a graph of the resulting data, the point at which the line
changes slope corresponds to the threshold energy, which is then related with the terminal
voltage and consequently used to calibrate the accelerator. The main advantage of this method
is that it can be used with a deuteron or a proton incident beam.
3.1.3. Elastic scattering
A third calibration method makes use of the elastic scattering of the ion beam [3.9, 3.10].
While this method can be used with any beam, it yields better results with a light ion beam. It
is less accurate, however, than the two above mentioned methods. The elastic scattering
method is based on the fact that ions produced by a two-body reaction are backscattered with
a well-defined energy due to kinematics. An ion beam is focused on a polished target and the
backscattered ions are detected with the aid of a Silicon Surface Barrier (SSB) or Silicon PIN
diode detector. Prior to the experiment, the detector should be calibrated using a point source.
The energy of the detected ions is then inserted into a set of equations describing two-body
kinematics, and the initial energy of the beam is calculated, which in turn corresponds to the
terminal voltage set. The whole procedure can then be repeated with different beam energies,
producing more points between the terminal voltage and the accelerator’s energy. While this
method has the advantage of being able to produce an infinite number of such points, its
accuracy is restricted by the resolution of the particle detector, which is usually around 15
keV, and also by the roughness of the target used.
3.2. EFFICIENCY CALIBRATION OF THE DETECTOR
The term efficiency, as it is used in Eq. 3.1, refers to the absolute efficiency of the detector
and not the intrinsic efficiency. When calculated with the procedure described below, it also
takes into account the solid angle covered by the detector.
Although efficiency calibration is a common procedure, it requires attention to detail in order

to produce accurate results. Depending on the energy of the gamma-ray, (i) a suitable point
source is placed as close as possible to the position of the actual target. The positioning of the
target is crucial since the solid angle must be taken into account as well as the attenuation of
the gamma-rays due to the materials between the source and the detector. The latter effect is
more important when low energy gamma-rays are concerned. (ii) The accumulated spectrum
is analysed and each gamma-ray peak area is corrected for the dead time of the measurement,
which should be kept as low as possible. (iii) The efficiency for each of the gamma-rays
emitted is calculated using:
Nγ
ε= , (3.1)
∆t ⋅ R ⋅ I
where Nγ represents the number of detected gamma-rays, R is the activity of the source, ∆t is
the time duration of the measurement and I is the branching ratio of the detected gamma-ray.
Next, (iv) efficiency ε is plotted against the energy of the gamma-ray and fitted with a suitable
function, taking into account the corresponding errors. There are a variety of functions that
have been proposed for fitting the experimental data [3.11, 3.12]. Finally, (v) the efficiency
39
for the given gamma-ray is calculated from the fitted function. Special care must be taken for
the calculation of the error of the detector’s efficiency at that energy. A common mistake is to
calculate the error by a quadratic summation of the errors of the fitting parameters, which is
mathematically incorrect since the fitting parameters are not independent variables, and also
leads to an overestimation of the error. Instead, the error must be estimated by calculating the
66% confidence band of the fitted function. It is worthwhile noting that while the various
proposed functions give approximately the same value for the efficiency in a region well-
defined by nearby gamma-rays, their confidence bands may deviate significantly. For this
reason, best practice is to use a source that emits gamma-rays close in energy to the one under
study.
If gamma-rays are recorded that are more energetic than the ones emitted by the source, there
are several possible methods for dealing with them. The simplest and least accurate method is
by extrapolating the fitted function obtained by a source with low gamma-rays. Because of
the intrinsic uncertainty that enters every extrapolation procedure, this method should be
avoided. A second more complicated but far more accurate method employs the gamma-rays
emitted by well-studied reactions, including well-known resonant reactions [3.13]. If the
cross-section or the strength of the resonance is known with great accuracy, the efficiency of
the detector can be deduced by solving the inverse problem. A third method, which is widely
used, is to simulate the detector’s efficiency by means of a simulation package such as
GEANT [3.14–3.17], or PENELOPE [3.18, 3.19]. After building the geometry of the detector,
the simulation should be checked against a low-energy gamma-ray source before being used
for the calculation of the efficiency at higher energies. The main advantage of this method is
that once the simulation is made for a specific detector, it can be used again for future
experiments using the same detector as long as the geometry of the experiment does not
change, or as long as the simulation is adjusted accordingly.
3.3. TARGET PREPARATION
The main requirement for all cross-section measurements is a thin target of known
composition. There are multiple ways of producing such a target, depending mainly on the
pure element or compound to be produced. The produced target may be self-supported or
deposited onto a thick backing. The backing should not emit additional gamma-rays. For this
reason, backings such as Ta, Cu or even Au with high purity are commonly used. A good
practice is to check the gamma-rays emitted from the backing before using it. This check can
also be performed at any point during the experiment by flipping the target so that the beam
faces the backing.
If the target can be obtained prefabricated from a commercial company, this is the
recommended option due to the guaranteed purity of the target. Independent measurements to
confirm the nominal thickness, however, should still be performed.
If the target cannot be obtained from a commercial company, it can be prepared in the lab
using specialized equipment such as evaporators. The simplest method uses thermal
evaporation to prepare targets of compounds with low evaporation points. The compound is
placed into a metal boat (Ta, Mo or W) inside a vacuum chamber. Current flows through the
boat heating the compound, which eventually evaporates. The vapour rises to deposit onto a
thick backing placed at some distance above the crucible boat. A second method uses an
electron gun evaporator within a vacuum chamber. A filament is heated to produce electrons,
under high vacuum, which are guided with the aid of magnets onto the compound. The
40
electrons sputter the atoms in the compound, which then deposit onto an appropriate backing
that has been placed inside the chamber. The third method, in the case of metallic targets,
involves rolling a foil several times until it reaches the desired thickness.
When performing any of the three mentioned manufacturing methods, there are a few key
considerations and precautions: (i) if the target is a poor thermal conductor, the heat produced
as a result of the beam hitting the target could change either the stoichiometry or the initial
thickness of the target. If excess heat occurs, it is advisable to use a low beam current and
duplicate some measurements in order to determine if any thermal damage has occurred. If
needed, the target should be cooled, however proper consideration should be taken when
deciding how to cool the target so the target is not damaged further and so that the integration
of the beam charge is not altered. (ii) There are different methods of measuring the beam
charge. In direct methods, the target must be a conductor so that no discharges will occur
during the experiment. Indirect methods detect backscattered particles and require that a thin
metal film of a heavy element, such as Au or Ag, be deposited onto the actual target. Further
discussion of this topic can be found in Section 3.4.
After a target is manufactured in the lab, it must be characterized. This can be achieved using
other ion beam analysis techniques such as RBS, NRA or Elastic Recoil Detection Analysis
(ERDA). The characterization is crucial not only because the thickness directly affects the
cross-section but also in order to discover any possible impurities in the target.
3.4. CHARGE COLLECTION
There are direct and indirect ways of measuring the charge [3.20]. The easiest direct way, if
the incident beam loses all of it’s’ energy within the target, is by electrically connecting the
target to a sensitive current integrator. If the target is thin enough, the beam may pass through
it, at which point it must be halted and collected in a Faraday cup. Although it seems a
straightforward and easy method, several potential sources of uncertainty and error must be
taken into consideration. The most important source of uncertainty in direct collection is the
production of secondary electrons. As the beam impinges on the target, it emits electrons that,
when allowed to escape the target, produce a ‘fake’ positive charge, with the amount of
electrons produced dependent on the nature of the target. The same effect can be produced if
electrons escape from the collimator and reach the target, which reduces the beam charge on
the target. A common way of resolving the problem is to apply a positive voltage on the target
and on the collimators so as to suppress the production of electrons. This suppression voltage
(typically ~ + 300 V), when used in conjunction with a Faraday cup surrounding the target, is
sufficient for a reliable charge collection. The setup must be checked, however, before each
experiment in order to ensure proper function.
A widely-used way of indirectly measuring the charge is by simultaneously detecting both the
gamma-rays and the beam particles scattered at backward angles. A SSB detector placed in a
backward angle between 150o and 170o detects the elastically backscattered ions coming from
a thin metallic film evaporated on the target. If the thickness of the film and the solid angle
covered by the SSB detector are known, the number of incident ions can be calculated with
the Rutherford formula. A heavy element for the film will ensure the validity of Rutherford’s
scattering law.
Besides the common method, other more elaborate indirect methods have been proposed. One
example of a different method is the use of a beam chopper, which is a device that interacts
41
with the beam in predefined time intervals and can directly and indirectly measure the beam
current.
3.5. CROSS-SECTION DETERMINATION
The differential cross-section is calculated using the following formula:
 dσ  N γ ( E ,θ )
 ( E,θ ) = , (3.2)
 dΩ  ξ ⋅ N p ⋅ ε αγ ( Eγ )
where Nγ(Ε,θ) is the number of detected gamma-rays, ξ is the thickness of the target in
atoms/cm2, Np is the number of incident particles and εαγ(Εγ) is the absolute efficiency of the
detector. Since the methods of determining quantities for most of the variables in Eq. 3.2 have
already been discussed in Section 2, only the method of determining Nγ is elaborated on here.
Depending on the multi-polarity of the gamma-ray, the emission can be anisotropic, which
requires the measurement of the cross-section at more than one angle, or isotropic. If there is a
possibility of an angular dependence, the angle should be defined with the best possible
accuracy acquired by placing the detector as far away from the target as possible. On the other
hand, a large distance dramatically reduces the counting rate of the experiment. To maximize
both counting rate and accuracy, a middle ground should be found. Once the distance has
been set, cross-sections can be measured either by simultaneously using multiple detectors
placed at different angles or by using one detector and making a new measurement at each
angle.
After setting up the detector and associated electronics, a test run should be performed to
study the gamma-rays coming from the background, either natural or from the interaction of
the beam with the building materials of the beamline. If there are no outside gamma-rays
emitted in the vicinity, the target can be placed in the chamber. If the target is backed, a test
run must be performed with the beam directed on the backing material to ensure that no
overlapping gamma-rays are emitted. It is recommended that the measurement be repeated at
a higher energy beam, since a reaction channel that was closed at a low energy may open at a
higher one.
Before starting the actual experiment, the energy step of the beam must be carefully selected.
This decision is based mainly on three factors: (i) the ripple of the accelerator, (ii) the
thickness of the target and (iii) the shape of the cross-section. While the first two can be taken
into account before starting the experiment, the third can only be determined during the
experiment, and then adjusted as needed, i.e. if sharp resonances exist in the cross-section.
One way of predicting any potential resonance energies is by studying the level scheme of the
compound nucleus, if it is available. During the experiment, the intensity of the incident beam
should be kept as low as possible so as not to damage the target or produce pile-up effects on
the detection system.
After recording the spectra, an appropriate integration or fitting computer code should be
used. In cases where a strong background exists, more than one computer code should be
used, at least for some measurements, so as to avoid any systematic errors in the integration
procedure. The measured cross-section can then be determined using Eq. 3.2. Finally, the
energy of the measured cross-section must be corrected using the accelerator’s calibration and
the beam’s energy loss within the target.
42
3.6. INTERLABORATORY EXERCISE: MEASUREMENT OF THE CROSS-SECTION
OF THE 27Al(p,p′t)27Al REACTION
After surveying the scientific literature for publications containing previous measurements of
cross-sections for PIGE spectroscopy, it was discovered that large discrepancies exist between
cross-sections of the same reaction measured by different authors in different laboratories. In
response, a concerted and coordinated effort was made to assess the experimental constraints
and difficulties associated with measurements of (p,p′γ) reactions.
An interlaboratory exercise was organized to measure the differential cross-section of the

27
Al(p,p′γ)27 Al reaction that produces the 844 keV gamma-ray. The emission of this gamma-
ray is isotropic due to the spin of the initial state which is Ji = 1/2. The measurements were
made at energies ranging from 2.5 MeV to 3 MeV with a 10 keV energy step. Additionally,
thick target yields were measured for the 27Al(p,γ)28Si reaction at energies between 0.98 and
1.1 MeV with the energy step small enough to reproduce the 991.9 keV resonance, using the
1.779 MeV gamma-ray line (see Table 3.1).
3.6.1. Implementation
The methodology used in the interlaboratory exercise follows the steps detailed in Sections
3.1–3.5. Each step of the measurement, including energy calibration, target preparation,
charge collection and efficiency determination, was discussed beforehand and guidelines were
agreed upon in order to minimize variance and uncertainty from different set-ups.
Unavoidable uncertainties associated with each step were estimated and collected for
comparisons into Table 3.4 for differential cross-sections and Table 3.5 for thick target yields.
In the following sections, each step of the interlaboratory exercise is presented and the results
are discussed.
Efficiency (ε): In order to minimize discrepancies related to the determination of gamma-ray

detector absolute efficiency εaγ, it was decided to use the branching ratios of calibration
source gamma-ray lines, as reported in a recent IAEA compilation [3.21], and fit the
experimental results by the following equation:
b c d
ε aγ (Eγ ) = a + + 2+ 3, (3.3)
Eγ Eγ Eγ
where Eγ is the energy of the gamma-ray line, and a, b, c, and d are fitting parameters.
Targets: For differential cross-section measurements, thin self-supporting Al films or thin

films of Al over thin self-supporting Ag (or Au) films were used. For thick target yield
measurements, freshly polished Al foils or freshly evaporated thick films of Al were used. In
this instance, ‘thin target’ means that the proton energy loss in the target, i.e. the energy
thickness of the target, is much smaller than the natural widths of the resonances to be
measured.
Cross-sections: Different approaches were used by the different laboratories in relation to

differential cross-section measurement and determination.
43
One approach is based on Eq. 3.2 for the differential cross-section at incident proton energy
E. Due to the isotropy of the emitted gamma radiation, the equation can be simplified to the
following:
dσ Aγ ( E )
(E) = , (3.4)
dΩ 4π ⋅ N p ⋅ N Al ⋅ ε aγ
where Aγ(E) is the peak area of the detected gamma-rays, Np and NAl are the number of
incident projectiles and the number of Al atoms per surface unit (target thickness ξ)
respectively, and εaγ is the absolute efficiency of the gamma-ray detector. To use this
equation, the composition of the target (usually determined by an independent RBS
measurement) and the absolute value of the collected charge related to each gamma-ray
spectrum must be known.
Since the above mentioned values are often difficult to accurately obtain, the following
approach is often used, which is based on simultaneously measuring the elastic scattering of
the projectiles from a heavy component of the target (which could be either a self-supporting
film or a very thin evaporated film). Using Eq. 3.4 and the yield of projectiles scattered from
the heavy element (H) in the target (Eq. 3.5), one can deduce Eq. 3.6, which is used for the
determination of the differential cross-section:
Asp ( E , β ) = σ R , H ( E , β )⋅ N p ⋅ N H ⋅ Ω p ⋅ ε ip , (3.5)
dσ Aγ ( E )⋅ σ R , H ( E , β )⋅Ω p ⋅ ε ip
(E ) = . (3.6)
dΩ 4π ⋅ r ⋅ Asp ( E , β )⋅ ε ip γ
In the above equations, Asp(E) is the peak area of the detected projectiles scattered from the
heavy element (H) at an angle β with respect to the beam direction, at a beam energy E.
σR,H(E,β) is the Rutherford cross-section at energy E and scattering angle β. Ωp is the solid
angle of the particle detector, with εip representing the intrinsic efficiency (typically equal to 1
for protons in the considered energy range). r is the ratio of the number of Al atoms to the
number of atoms of the heavy element.
When using the normalization procedure of Eq. 3.6, the underlying assumption is that the
elastic scattering cross-sections for heavy elements at the energies applied are purely
Rutherford. In this method, the uncertainty related to the determination of the absolute value
of the collected charge, which could be very large [3.20], can be avoided by using the
Rutherford cross-section for heavy elements, which is precisely and accurately described.
Another advantage of this method is that the thickness of the target is not required: only
elemental ratios need to be known. Since the elemental ratios of the compound can be
determined with a higher degree of accuracy, the uncertainty of the calculation is further
minimized. The cross-section determination using Eq. 3.6 may be done for each measured
energy point or for just one (plus a few for cross-checking), with the rest of the measurement
points normalized to this point. This implies relying upon the reproducibility of relative
charge measurements, with corresponding uncertainty usually much smaller than the one
affecting the absolute value of the collected charge.
A third approach is based on the transformation of Eq. 3.4 into Eq. 3.7:
44
dσ Aγ (E)⋅ S AA
(E) = , (3.7)
dΩ 4π ⋅ N p ⋅ ∆ E ⋅ ε aγ
where ∆E is the energy thickness of the target and SAA is the stopping cross-section in units of
energy by atoms per area squared. The target energy thickness may be measured by RBS or
determined from the experimental width of a resonance with very small natural width, e.g. the
one at 991.9 keV in the 27Al(p,γ)28Si reaction. In this approach, care must be taken when
determining SAA with due consideration of the possible oxidation of the aluminium film. In
this particular interlaboratory exercise, some of the laboratories followed the approach related
to Eq. 3.4, while others determined their cross-sections by Eq. 3.6 or Eq. 3.7.
For thick target measurements, the gamma-ray yield Yγ(E) as a function of proton beam
energy E can be calculated using the following formula:
Aγ (E)
Yγ (E) = , (3.8)
ε aγ ⋅Q ⋅ fi ⋅ fm ⋅ N Av ⋅ A−1
where Q is the integrated proton beam current, fi is the abundance of the isotope producing
the gamma-radiation, fm is the mass fraction of the analysed element, NAv is the Avogrado
number and A is the atomic mass of the element. Since pure Al targets were used, both fi and
fm are equal to 1 in this case.
3.6.2. Experiment
The experimental conditions employed in each laboratory are described in full in Section 4.
The essential experimental details related particularly to this interlaboratory exercise are
summarized in Table 3.2 and Table 3.3.
45
TABLE 3.2. EXPERIMENTAL DETAILS FOR EACH OF THE PARTICIPATING LABORATORIES (Five out of ten laboratories)
NCSR “Demokritos” CTN/ IST KIPT NSTRI RBI
Terminal voltage X
Magnetic field X X X X
Calibration points 992 keV, 992 keV, 992 keV, 992 keV, 992 keV,
27 27 27 27 27
and methods Al(p, γ) 28Si Al(p,γ) 28Si Al(p,γ) 28Si Al(p,γ) 28Si Al(p,γ) 28Si
Peak position in RBS 672, 872 and 1371 keV, Peak position in 1880 keV 1880 keV threshold,
Calibration of 19 7
spectra F(p,αγ)16O gamma-ray threshold, Li(p,n)7Be
projectile energy 7
1645 and 1931 keV, spectra (direct Li(p,n)7Be
23
Na(p,p′γ)23Na capture)
2400, 2914 and
3660 keV,
24
Mg(p,p′γ)24Mg
Type HPGe HPGe DGDK 80B HPGe HPGe
Active Volume ~80 cm3 213 cm3 108 cm3
Distance to target 15 cm 50 mm 50 mm 51.9 mm 66 mm
Angle to beam 90° 130° 90° 90° 135°
Detector parameters direction
Absolute efficiency 1.5 × 10-3 3.2 × 10-3 21 × 10-3 11.9 × 10-3 3.0 × 10-3
at 844 keV
Absolute efficiency — 2.1 × 10-3 9.1 × 10-3 5.7 × 10-3 1.4 × 10-3
at 1779 keV
Material and Pure Al (equivalent 28 µg/cm2 Al on — 26 µg/cm2 Al on 55 µg/cm2 Al on
Target for thickness thickness 3 keV for 100 µg/cm2 Ag 129 µg/cm2 Ag 6.3 µm Mylar,
differential cross- 992 keV protons) 3.5 µg/cm2 Au top
section layer
measurements Thickness Gamma-ray resonance RBS with alphas — RBS with alphas RBS with alphas
determined by
46
TABLE 3.2. EXPERIMENTAL DETAILS FOR EACH OF THE PARTICIPATING LABORATORIES (Five out of ten laboratories) (cont.)
NCSR “Demokritos” CTN/ IST KIPT NSTRI RBI
Material and — Thick Al foil Thick Al foil — —
thickness
Charge collection Target and chamber as Target and chamber as a Target and Faraday cup Target and Faraday
Target for thick
a Faraday cup Faraday cup chamber as a after the target cup after target,
target gamma-ray
Faraday cup suppressor around
yield measurements
target
Normalization Eq. 3.6 Eq. 3.7 Eq. 3.7 Eq. 3.7
procedure
47
TABLE 3.3. EXPERIMENTAL DETAILS FOR EACH OF THE PARTICIPATING LABORATORIES (other five laboratories)
ATOMKI RUB CMAM-UAM Helsinki LABEC- INFN
Terminal voltage X X
Magnetic field X X X
Calibration points 992 keV, 342 keV , 992 keV, 3903, 3996 and 992 keV,
27 19 27 27
and methods Al(p,γ) 28Si F(p,α γ)16O Al(p, γ) 28Si 5937 keV, Al(p, γ) 28Si
14
992 keV, Non resonant N(p,p′ γ)14N 1683 keV,
27 27
Al(p, γ) 28Si particle emitting Al(p,p′ γ)27Al
Calibration of
1666 keV, nuclear reaction and 2876 keV,
projectile energy 4 27
He(11B, γ)15N (*) RBS according to Al(p,p)27Al
5663 keV, INDC(NDS)-0589 3470 keV,
4 16
He(14N, γ)18F (*) p. 30 O(p,p)16O
( )
* inverse
kinematics with gas
target
Type HPGe HPGe Reverse Electrode Coaxial Ge HPGe
Ge
Active Volume 170 cm3 110 % rel. to 178 cm3 38% rel. to 4” × 4” 235 cm3
4” × 4” NaI(Tl) NaI(Tl)
Distance to target 95 mm 135 mm 210.6 mm 20 mm 205 mm
Detector parameters Angle to beam 55° 90° 135° 55° 90°
direction
Absolute efficiency 3.6 × 10-3 3.2 × 10-3 0.9 × 10-3 3.1 × 10-3 1.2 × 10-3
at 844 keV
Absolute efficiency 1.9 × 10-3 2.1 × 10-3 0.6 × 10-3 1.5 × 10-3 0.8 × 10-3
at 1779 keV
Material and 39 µg/cm2 Al on Pure Al (equivalent 22 µg/cm2 Al on 203 µg/cm2 Al 29 µg/cm2 Al on
Target for thickness 33 µg/cm2 C thickness 6.9 keV 39 µg/cm2 Ag 102 µg/cm2 Ag
differential cross- for 992 keV
section protons)
measurements Thickness RBS with alphas Gamma-ray RBS with Li ERDA RBS with alphas;
determined by resonance PIXE with protons
48
TABLE 3.3. EXPERIMENTAL DETAILS FOR EACH OF THE PARTICIPATING LABORATORIES (other five laboratories) (cont.)
ATOMKI RUB CMAM-UAM Helsinki LABEC- INFN
Material and — Pure Al (equivalent — — Thick Al foil
thickness thickness 6.9 keV
for 992 keV
Target for thick
protons)
target gamma-ray
Charge collection Target and chamber Polarized target Target and chamber Faraday cup after Target and Faraday
yield measurements
as a Faraday cup as a Faraday cup target cup after the target
Normalization Eq. 3.4 Eq. 3.4 Eq. 3.6 Eq. 3.4 Eq. 3.7
procedure
49
3.6.3. Results and discussion
3.6.3.1. Differential cross-section measurements
The results of the differential cross-sections obtained from the reaction 27Al(p,p′γ)27Al, Eγ =
844 keV by the nine laboratories that participated in this interlaboratory exercise (ATOMKI
[3.22], University of Bochum [3.23], CMAM-UAM [3.24], University of Helsinki [3.25],
LABEC-INFN [3.26], NCSR “Demokritos” [3.27], CTN/IST [3.28], NSTRI [3.29] and RBI
[3.30]) are displayed in the first rows of Table 3.2 and Table 3.3 under the row heading
‘Calibration of projectile energy’. The estimated uncertainty budget for the measurement
parameters that contribute to the absolute differential cross-section values, for each
laboratory, is reported in Table 3.4. The large discrepancies in cross-section values and proton
energy calibration, which are the results of the uncertainties present within each experimental
set-up, are illustrated in panel II of Fig. 3.2. These discrepancies introduce another degree of
difficulty to the accurate assessment of cross-section values. Two main reasons why the
resonance points from different laboratories are not directly comparable are (i) the large
energy steps that were used, leading to poorly-defined resonances, and (ii) energy calibration
discrepancies, which led to different laboratories measuring different energy points.
To assist in the assessment of the experimental values, an evaluated cross-section was

produced [3.31]. As part of the evaluation process, a theoretical gamma-ray production cross-
section for the 27Al(p,p′γ)27Al reaction was obtained from the calculation of the inelastic (p,p′)
cross-section in the framework of the R-matrix theory [3.32]. According to [3.31], in the
energy range under consideration proton inelastic scattering by Al leaves the residual nucleus
mainly in the first excited state (Jπ = ½+, Ex = 844 keV), with ~3% contribution from the
transition from the second excited state (Jπ = 3/2+, Ex = 1015 keV). Both the cross-sections for
the 27Al(p,p1)27Al and 27Al(p,p2)27Al reactions were calculated and the contribution from the
second excited state could be taken into account. The initial data for the resonance parameters
that were required for the R-matrix calculations were taken from [3.33]. Comparisons were
then made between the theoretical cross-sections and all the experimental data from the
interlaboratory exercise, taking into account the uncertainty budget. Based on this analysis, an
evaluated cross-section was produced, which represents the best estimation for the actual
cross-section value. This evaluated cross-section is illustrated as a solid line in Figs. 3.2 and
3.3.
50
TABLE 3.4. ESTIMATED UNCERTAINTY BUDGET FOR MEASUREMENT PARAMETERS THAT CONTRIBUTE TO THE ABSOLUTE
DIFFERENTIAL CROSS-SECTION VALUES GENERATED BY EACH OF THE LABORATORIES (ERROR: 1 STANDARD DEVIATION)
Parameters Uncertainty
ATOMKI RUB CMAM- Helsinki LABEC- NCSR CTN/IST NSTRI RBI
UAM INFN “Demokritos”
γ-ray peak area (Aγ )a 1.0% 1.0–4% 1% 1.5–3% 1.0–3% 1–3% 1.0–3% 1.0% 1.0–
4%
p+ heavy element (H) elastic peak area ( Asp) 1.0% 1.0% 1.0% < 1%
p+ heavy element (H) Rutherford cross-section 1.0% 1.0% 1.0% 1.0%
(σR,H )b
Al-to-heavy element (H) atomic ratio (r) 3% 3% 4% 2.0%
Al atomic density (NAl) 3% 5% 3%
energy thickness of the target (∆E) 1.5% 2%
stopping power (SAA) 4% 4%
number of incident protons (Np) 3% 10% 1–2% 3% 5%
particle detector solid angle (Ωp) 2.4% 2.5% 2.8% 2.0%
absolute γ-ray efficiency (εαγ) 4% 10% 4% 8% 6% 5% 6% 7% 7%
proton beam energy (E) 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% < 0.1% 0.1%
a
When two values are indicated, the first refers to resonant and the second to off-resonance energy points, respectively.
b
Uncertainty originates in the proton beam energy in the heavy element layer in the target and in the particle detector scattering angle.
51
(I)
(II)
FIG. 3.2. Panel I displays differential gamma-ray production cross-sections of the 844 keV gamma-
ray of the 27Al(p,p′γ)27Al reaction as a function of proton energy, deduced from work of each of the
nine laboratories participating in this exercise: a) [3.22], b) [3.26], c) [[3.30], d) [3.25], e) [3.24],
f) [3.28], g) [3.23], h) [3.29] and i) [3.27]. The solid line ( j) [3.31]) corresponds to the evaluated
cross-section. Panel II is a close up view of panel I. Within the resonance (peak) region, the large
deviations in proton energy between the different datasets are more noticeable.
The energy scales of the different laboratories were then adjusted with respect to the
evaluated cross-section, by adding or subtracting amounts up to 8 keV, but less than 3–σ of
the assessed uncertainty on the calibrated proton beam energy. Using the same energy
adjustment for the entire energy interval allows a better comparison of the excitation curves,
as shown in Fig. 3.3.I. The discrepancies between measurements made at different
laboratories are most apparent in the resonance regions, as shown in Fig. 3.2.II. Evident in Fig
3.3.II, which is a close up of Fig. 3.3.I, is that data outside the resonance region more closely
agree and, furthermore, most of the measurements agree within the quoted total uncertainties
(10–12%), which stem from the measurement parameters (Table 3.4).
52
(I)
(II)
FIG. 3.3. Panel I displays differential gamma-ray production cross-sections of the 844 keV gamma-
ray of the 27Al(p,p′γ)27Al reaction as a function of proton energy, after the application of a correction,
which is based on the evaluated cross-section, to the energy scales reported by the different
laboratories: a) [3.22], b) [3.26], c) [3.30], d) [3.25], e) [3.24], f) [3.28], g) [3.23], h) [3.29] and
i) [3.27]. The solid line corresponds to the evaluated data: j) [3.31]. In panel II, which is a close up of
panel I, it is shown that, outside of the narrow resonance regions, most of the measurements agree
within the quoted total uncertainties (10–12%).
After evaluating the approach used by each laboratory, it can be concluded that the results
obtained through methods that were not directly dependent on the absolute value of the
collected charge are in slightly better agreement and have discrepancies closer to the quoted
uncertainties.
Figure 3.2.II shows in larger detail the discrepancies in projectile energy calibration, raising
questions concerning the cause of such large deviations. One of the difficulties with
calibrating, particularly for energies higher than 2 MeV, is that there are few narrow
resonances to rely on. Table 3.2 and Table 3.3 list the resonances that have been employed for
53
projectile energy calibration. Since there is a lack of consistency of the energy values of those
resonances in the literature, this may be one cause of the discrepancies in projectile energy
calibration.
As an example, in the CTN/IST laboratory, proton energy calibration was first performed
using the high energy resonances of the 24Mg(p,p′γ)24Mg reaction, which were measured
earlier by Boni et al. [3.34]. The calibration was later repeated with the resonance for
16
O(p,p)16O at 3470 keV. The fit using the 16O resonance instead of the 24Mg resonances that
have energies higher than 2 MeV, leads to a slope that implies a better energy scale agreement
between the measured data points of the CTN/IST laboratory and the evaluated cross-section.
Support for the discrepancies of the 24Mg resonance energies in comparison with the 25Al
resonance energies can be found in recent literature. The energies quoted in the paper of Boni
et al. are in close agreement with excited states of 25 Al taken from the compilation of Endt
and van dèr Leun [3.35]. However recent experimental work [3.36] assigns higher energy
values to excited states of 25Al above 5 MeV, the region of the 2914 and 3660 keV
resonances.
Regarding the cross-section values, discrepancies higher than 3 times the quoted uncertainties
(7–12% at 1σ level) may only be explained by systematic deviations in the assessment of the
target elemental ratios and/or of either (a) the particle detector solid angle, if the approach
based on Eq. 3.6 is followed, or of (b) the charge and target thickness, if the approach based
on Eq. 3.4 or 3.7 is applied. Although background and multiple scattering may give a more
than negligible contribution to the particle spectra, such spectra are simple to deal with if the
following conditions are met: (i) scattered protons originating near the chamber walls and
components do not reach the particle detector and (ii) the heavy element (i.e Ag or Au) peak
within the scattered proton spectra is completely separated from other peaks, allowing easy
quantification. Moreover, a non-negligible background peak around 844 keV due to
interaction of the proton beam with aluminium, stainless steel (847 keV gamma-ray line from
the 56Fe(p,p′γ)56Fe reaction) or other materials containing Al or Fe, even at trace levels, found
either along the beamline or in the scattering chamber (collimators, beam stopper, Faraday
cup) could contribute to the production of higher cross-section values in the off-resonance
regions. Integration of all the resonances over the total target thickness smooths and broadens
the resonance shape so that the measured cross-section values appear lower, especially in
thick targets.
3.6.3.2. Thick target yield measurements
In determining the yield of the gamma-ray with energy Eγ = 1779 keV from the 991.9 keV
resonance of the 27Al(p,γ)28Si reaction, several thick or semi-thick target measurements were
made (University of Bochum [3.23], LABEC-INFN [3.37], CTN/IST [3.28] and KIPT
[3.38]). The estimated uncertainty budget associated with the measured absolute thick target
gamma-ray yields is reported in Table 3.5.
54
TABLE 3.5. ESTIMATED UNCERTAINTY BUDGET FOR THE MEASURED THICK TARGET
GAMMA-RAY YIELDS PERFORMED IN DIFFERENT LABORATORIES (ERROR: ONE
STANDARD DEVIATION)
Quantity Uncertainty
RUB LABEC- INFN CTN/ IST KIPT
γ-ray peak area 3% 5–8% 4% 
charge integration (absolute 10% 5% 8% 
value)
absolute γ-ray efficiency 10% 6% 5% 
proton beam energy 0.1% 0.1% 0.1% 0.1%
In Fig. 3.4, a thick target measurement is shown, whereas in Fig. 3.5 a semi-thick one is
displayed. Here, ‘semi-thick’ does not mean ‘thin’, as defined earlier, but rather that the target
has a thickness smaller than the proton beam energy range. The smooth line is a generated fit
to the data based on the integral of a Breit-Wigner resonance shape with a constant
background term, which takes into account the contribution to the integrated yield of
resonances at lower energy:
 ( E − b) π 
Yγ (E) = a ⋅ arctan 2 ⋅ + + d . (3.9)
 c 2
FIG. 3.4. Thick target gamma-ray yield of the 1779 keV gamma-ray from the 27Al(p,γ)28Si reaction as
a function of proton energy, around the 991.9 keV resonance. Function deduced from the experimental
data of a) [3.37]. Only statistical errors are shown. The arrow indicates the measured resonance
energy.
55
FIG. 3.5. Semi-thick target gamma-ray yield of the 1779 keV gamma-ray from the 27Al(p,γ)28Si
reaction as a function of proton energy, around the 991.9 keV resonance, as deduced from the
experimental data of a) [3.23]. Only statistical errors are shown. The arrow indicates the measured
resonance energy.
Without taking into account information on the energy resolution, i.e. assuming a fresh
polished smooth surface or a freshly evaporated Al film, the maximum yield for a thick target
Ymax(∞) (corresponding to a – π with a obtained from the fitting function of Eq. 3.9) divided
by the number of incident particles and the absolute efficiency, should be constant at the same
detection angle (the yield is weakly dependent on the detection angle [3.13]) and given by
[3.39]:
Ymax (∞) λ2 M +m 1
= ωγ (3.10)
N pε aγ 2 M S AA
where λ is the centre-of-mass de Broglie wavelength of the projectile, ωγ is the resonance

strength, m and M are the masses of the projectile and target, respectively. The mass ratio
corrects for the energy term of the stopping cross-section, which is measured or calculated in
the laboratory coordinate system. The yields given in Table 3.6 correspond to the first term of
Eq. 3.10 divided by 4π. Correction factors were applied to obtain the thick target yields from
semi-thick target yields.
TABLE 3.6. YIELDS FOR RELEVANT PARAMETERS RELATED TO THE MEASUREMENT

OF THE 991.9 keV RESONANCE OF THE 27Al(p,γ)28Si REACTION
Parameters RUB LABEC- INFN CTN/ IST KIPT
Maximum yield (Nγ/µC/sr) 617 ± 90 550 ± 55 610 ± 60* 600 ± 50
Resonance energy (keV) 993 ± 1 991 ± 1 992 ± 1 991 ± 1
Energy spread (keV) 1 2 2 4
56
The maximum yield for CTN/ IST was originally measured to be 618 Nγ/µC/sr at an angle of
130°, but was multiplied by an angular distribution correction factor [3.13] so that the value
could be applied to a 90° angle in order to compare it with the values from the other labs. The
results are in agreement with each other within 10% uncertainty, and the weighted average
and its associated error give ωγ = 2.01 ± 0.11 eV, which is in agreement within the quoted
uncertainty with the reference value of 1.93 ± 0.13 eV [3.22] .
3.6.3.3. Conclusions
The goal for the interlaboratory exercise was to assess and evaluate different methods for the
experimental determination of cross-sections and thick target yields of gamma-ray producing
nuclear reactions in order to understand the large discrepancies that exist among cross-section
data and to determine methods to minimize the discrepancies. Common methodology and
conditions were established at the beginning and two simple cases were chosen: one using a
thin Al target and one using a thick Al target. Results indicate that one of the major sources of
uncertainty is the accelerator energy calibration. Relative differences among laboratories were
greater than 10 keV. This may be caused by the lack of consistency in the existing data on
resonance energies used for accelerator energy calibration in the energy region between 2 and
4 MeV. In regards to the differential cross-section measurements, the results obtained from
laboratories that used a method not directly dependent on the absolute value of the collected
charge were found to be in slightly better agreement, with discrepancies closer to the quoted
uncertainties.

Analysis, 2nd edn, Materials Research Society, Pennsylvania (2009) p. 147.
[3.2] RÄISÄNEN, J., TIKKANEN, P.O., Calibration of accelerator proton energy by the
14
N(p,p´γ)14N reaction resonances, Nucl. Instrum. Methods Phys. Res. A 723 (2013)
5.
ROLFS, C., SCHULTE, W.H., BECKER, H.W., The 27Al(p,γ)28Si reaction: direct
capture cross- section and resonance strengths at Ep = 0.2–1.12 MeV, Eur. Phys. J. A
9 (2000) 479.
[3.4] MARION, J.B., HAGEDORN, F.B., 1.75 Mev resonance in C13(p, γ)N14, Phys. Rev.
104 (1956) 1028.
[3.5] RAO, P., KUMAR, S., VIKRAMKUMAR, S., RAJU, V.S., Measurement of
differential cross-sections and widths of resonances in 32S(p,p’γ) 32S reaction in the
3.0 – 4.0 MeV region, Nucl. Instrum. Methods Phys. Res. B 269 (2011) 2557.
[3.6] FREEMAN, J.M., (p,n) threshold measurements for accelerator calibration and the
problem of Q-value deductions, Nucl. Instrum. Methods 134 (1976) 153.
[3.7] SHAMU, R.E., BERNSTEIN, E.M., PARROTT, M.J., Accelerator energy
calibration using the 12 C(α,n0)15O threshold, Nucl. Instrum. Methods 114 3 (1974)
605.
[3.8] OVERLEY, J.C., PARKER, P.D., BROMLEY, D.A., The energy calibration of
tandem accelerators, Nucl. Instrum. Methods 68 (1969) 61.
[3.9] FERRERO, A.M.J., GARCIA, A., GIL, S., ETCHEGOYEN, A., DI TADA, M.,
PACHECO, A.J., ABRIOLA, D., DIGREGORIO, D.E., ELGUE, M.,
ETCHEGOYEN, M.C., FERNÁNDEZ NIELLO, J.O., MACCHIAVELLI, A.O.,
57
TESTONI, J.E., Improved beam-energy calibration technique for heavy-ion
accelerators, Nucl. Instrum. Methods Phys. Res. B 42 3 (1998) 389.
[3.10] SANTRA, S., MAHATA, K., SINGH, P., FERNANDES, C.V., HEMALATHA, M.,
KAILAS, S., Beam energy calibration for the folded tandem ion accelerator, Nucl.
Instrum. Methods Phys. Res. A 496 (2003) 44.
[3.11] DEBERTIN, K., HELMER, R.G., Gamma and X-Ray Spectrometry with
Semiconductor Detectors, North Holland, Oxford (1988).
[3.12] KISS, Z., FAZEKAS, B., OSTOR, J., REVAY, ZS., BELGYA, T., MOLNAR, G.L.,
KOLTAY, L., Comparison of efficiency functions for Ge gamma-ray detectors in a
wide energy range, Nucl. Instrum. Methods Phys. Res. A 418 (1998) 374.
[3.13] ANTTILA, A., KEINONEN, J., HAUTALA, M., FORSBLOM, I., Use of the 27Al(p,
y)28Si, Ep = 992 keV resonance as a gamma-ray intensity standard, Nucl. Instrum.
Methods 147 (1977) 501.
[3.14] KARAMANIS, D., LACOSTE, V., ANDRIAMONJE, S., BARREAU, G., PETIT,
M., Experimental and simulated efficiency of a HPGe detector with point-like and
extended sources, Nucl. Instrum. Methods Phys. Res. A 487 (2002) 477.
[3.15] ANIL KUMAR, G., MAZUMDAR, I., GOTHE, D.A., Experimental measurements
and GEANT4 simulations for a comparative study of efficiencies of LaBr3:Ce,
NaI(Tl), and BaF2, Nucl. Instrum. Methods Phys. Res. A 610 (2009) 522.
[3.16] AGOSTINELLI, S., ALLISON, J., AMAKO, K., APOSTOLAKIS, J., ARAUJO, H.,
et al., Geant4—a simulation toolkit, Nucl. Instrum. Methods Phys. Res. A 506
(2003) 250.
[3.17] ALLISON, J., AMAKO, K., APOSTOLAKIS, J., ARAUJO, H., ARCE DUBOIS,
P., et al., Geant4 developments and applications, IEEE Trans. Nucl. Sci. 53 (2006)
270.
[3.18] PEYRES, V., GARCIÁ-TORANO, E., Efficiency calibration of an extended- range
Ge detector by a detailed Monte Carlo simulation, Nucl. Instrum. Methods Phys.
Res. A 580 (2007) 296.
[3.19] JURADO VARGAS, M., GUERRA, A.L., Application of PENELOPE code to the
efficiency calibration of coaxial germanium detectors (Proc. 15th Int. Conf.
Radionucleotide Metrology and its Applications Oxford, 2005), Applied Radiation
and Isotopes 64 10–11 (2006) 1319.
Analysis, 2nd edn, Materials Research Society, Pennsylvania (2009) p. 347.
[3.21] INTERNATIONAL ATOMIC ENERGY AGENCY, “Recommended decay data,
high energy gamma ray standards and angular correlation coefficients”, Update of X
Ray and Gamma Ray Decay Data Standards for Detector Calibration and Other
Applications, Vol. 1, IAEA, Vienna (2007).
[3.22] CSEDREKI, L., et al., unpublished data, available in IBANDL [https://www-
nds.iaea.org/exfor/ibandl.htm].
[3.23] BECKER, H.-W., unpublished data.
[3.24] ZUCCHIATTI, A., et al., unpublished data.
[3.25] RAISANEN, J., unpublished data, available in
[https://www-nds.iaea.org/exfor/ibandl.htm].
[3.26] CHIARI, M., MELON, B., SALVESTRINI, L., FONSECA, M., ALVES, E.,
JESUS, A.P., Measurement of proton induced γ-ray emission cross-sections on Al
from 2.5 to 4.1 MeV, Nucl. Instrum. Methods Phys. Res. B 332 (2014) 355.
[3.27] LAGOYANNIS, A., et al., unpublished data.
[3.28] JESUS, A.P., et al., unpublished data, available in IBANDL [https://www-
nds.iaea.org/exfor/ibandl.htm]
58
[3.29] JOKAR, A., KAKUEE, O., LAMEHI-RACHTI, M., SHARIFZADEH, N.,
FATHOLLAHI, V., Differential cross-section measurements of 27Al(p,p’γ)27Al and
27
Al(p,αγ)24Mg reactions in the energy range of 1.6–3.0 MeV, Nucl. Instrum.
[3.30] ZAMBONI, I., et al., unpublished data, available in IBANDL [ https://www-
[3.31] GURBICH, A.F., SigmaCalc recent development and present status of the evaluated
cross-sections for IBA, Nucl. Instrum. Methods Phys. Res. B 371 (2015) 27.
[3.32] LANE, A.M., THOMAS, R.G., R-matrix theory of nuclear reactions, Rev. Mod.
Phys. 30 (1958) 257.
[3.33] NELSON, R.O., BILPUCH, E.G., WESTERFELDT, C.R., MICHELL, G.E., Proton
resonances in 28Si from Ex =13.4 to 14.5 MeV, Phys. Rev. C 30 (1984) 755.
[3.34] BONI, C., CEREDA, E., BRAGA MARCAZZAN, G.M., DE TOMASI, V.,
Prompt gamma emission excitation functions for PIGE analysis, Nucl. Instrum.
[3.35] ENDT, P.M., VAN DÈR LEUN, C., Energy levels of A = 21–44 nuclei (V),
Nucl.Phys. A 214 (1973) 1.
[3.36] PITTMAN, S.T., BARDAYAN, D.W., BLACKMON, J.C., KOZUB, R.L., SMITH,
M.S., Analysis of 25Al energy levels observed in the 28Si (p,α) 25Al reaction, Phys.
Rev. C 80 (2009) 038801.
[3.37] CHIARI, M., unpublished data, available in IBANDL [https://www-
[3.38] GONCHAROV, A., et al., unpublished data.
[3.39] ROLFS, C.E., AND RODNEY, W.S., Cauldrons in the Cosmos: Nuclear
Astrophysics, University of Chicago Press, USA (1988).
59
4. MEASUREMENTS OF PARTICLE-INDUCED GAMMAR-RAY EMISSION
(PIGE) CROSS-SECTIONS
One of the most important goals of the CRP was to improve the existing experimental cross-
section database for Particle Induced Gamma-ray Emission (PIGE) by performing new cross-
section measurements. Measurements were especially important in cases where no data
existed or where unresolved discrepancies between measured data needed clarification. At the
beginning of the CRP, it was important to unify the approaches used by participants from
different laboratories in order to minimize potential discrepencies and errors in cross-section
measurements. The methodology, described in detail in Section 3, was adopted by all CRP
participants and an effort was made to provide, apart from the absolute values of the cross-
sections, a well documented uncertainty budget that considered all the known factors (for
details see Section 3) including: the counting statistics of peak areas, thickness of the target,
current integration, γ-ray detector absolute efficiency, detector solid angle and dead time.
When possible, thin targets were selected to avoid problems related with cross-sections
calculated from thick target yields. The acccelerator energy was calibrated using well known
resonant reactions, such as 27Al(p,γ)28Si at 992 keV, 19F(p,αγ)16O at 872 keV and neutron
threshold reactions such as 7 Li(p,n)7Be at 1880.6 keV. Uncertainties in the detector angular
settings were minimized to a negligible value. When possible, normalization to a Rutherford
cross-section was done in order to remove uncertainties due to dead time, improper charge
collection and solid angle. The energy step for cross-section measurements was adjusted
according to the complexity of the excitation function in order to record all the necessary
details.
In the rest of Section 4, all measurements that were completed within the framework of the
CRP are explained in detail. All energies, angles and cross-sections are given in the laboratory
system. Performed measurements are organized according to the target element. New
measurements were compared with the data already existing in the literature, and a critical
assessment of all data was performed. Where possible, recommendations concerning which
data sets to use are stated. Obtained excitation functions are given in graphical form.
Tabulated values were uploaded onto IBANDL (http://www-nds.iaea.org/ibandl/). Measured
cross-sections are presented in units (mb/sr). When the angular distribution of the emitted γ-
ray is isotropic, the units (/sr) reflect the solid angle of the detector. When the angular
distribution is anisotropic, the angles of detection are explicitly mentioned in the graphs. Most
of the data were published in scientific journals and a complete set of references is given at
the end of Section 4.
The laboratories that participated in the CRP measurements:
— Centro de Micro Análisis de Materiales, Universidad Autónoma de Madrid, (CMAM-

UAM), Madrid
— University of Helsinki (Helsinki), Helsinki
— Institute for Nuclear Research, Hungarian Academy of Sciences (ATOMKI), Debrecen
— Kharkov Institute of Physics and Technology (KIPT), Kharkov
— Campus Tecnologico Nuclear/Institut Superior Tecnico (CTN/IST), Lisbon
— LABEC – Istituto nazionale di Fisica Nucleare (LABEC-INFN), Florence
61
— National Centre for Scientific Research “Demokritos” (NCSR “Demokritos”), Athens
— Nuclear Science and Technology Research Institute (NSTRI), Tehran
— Ruđer Bošković Institute (RBI), Zagreb
— Ruhr Universität Bochum (RUB), Bochum
4.1. LITHIUM
4.1.1. Gamma-ray production cross-cections of lithium from proton-induced nuclear

reactions
Previous thick-target yield measurements on natural lithium [4.1–4.3] determined that there
are two γ-rays that have a large enough yield for analytical purposes: the 429 keV γ-ray from
the 7Li(p,nγ1–0)7Be reaction and the 478 keV γ-ray from the 7 Li(p,pγ1–0)7Li reaction. Both γ-
rays are expected to have an exactly isotropic angular distribution. The (p,γ) radiative capture
reactions of both lithium isotopes (6Li, 7.5% abundance, and 7Li, 92.5% abundance) have just
astrophysical significance.
4.1.1.1. 7Li(p,nγ1–0)7Be: 429 keV γ-ray
Apart from the data already included in the IBANDL compilation, only the paper from Rio et
al. [4.4] presents emission yields (Fig. 1d from [4.4]) of the 429 keV γ-ray from the 7Li(p,nγ1–
7
0) Be reaction for proton energies from 1 to 3 MeV. However, since the units used in the
paper (counts per µC) are arbitrary, the data are of limited practical use for the PIGE
technique. Table 4.1 reports the only two cross-section data sets found in the literature and
Table 4.2 reports the only two thick-target yield data sets.
TABLE 4.1. DATA IN THE LITERATURE FOR THE DIFFERENTIAL GAMMA-RAY

PRODUCTION CROSS-SECTIONS OF 429 keV 7Li(p,nγ1–0)7Be
Reference Data Angle Ep (MeV) Target Quoted Data
source uncertainties presentation
[4.5] IBANDL 90° 2.2–3.8 1000 Å LiF 15% Tabular (R33
evaporated on file)
Nuclepore filters
[4.6] IBANDL 135° 3.0–5.5 50 µg/cm2 LiF 1–20% Tabular (R33

on 30 mg/cm2 C, statistical, file)
coated with 13.5%
20 µg/cm2 Au systematic
62
TABLE 4.2. DATA IN THE LITERATURE FOR THE THICK-TARGET YIELD OF 429 keV
7
Li(p,nγ1–0)7Be
Reference Data Angle Ep (MeV) Target Quoted Data
[4.2] IBANDL 55° 3.1–4.2 Pellet of Li — Tabular (R33
chemical file)
compound
[4.3] IBANDL 55° 7.0–9.0 Pellet of Li 2–5% Tabular (R33

powdered statistical, file)
chemical 7% systematic
compound,
1 mm thick
Figure 4.1 portrays the cross-sections listed in Table 4.1. The proton energy and the
differential cross-sections are given in the laboratory frame of reference, with energy units in
keV and cross-section units in mb/sr.
FIG. 4.1. Differential γ-ray production cross-sections of the 429 keV 7Li(p,nγ1–0)7Be resonance as a
function of proton energy, deduced from the works of a) [4.6] and b) [4.5]. Only statistical errors are
shown.
The two cross-section data sets in Fig. 4.1 agree in shape, but data from Boni et al. [4.5] are
about 30–35% higher than data from Caciolli et al. [4.6]. The quoted total uncertainties for
the two data sets are 15% and 13.5% respectively. To solve this discrepancy and determine a
recommended cross-section data set, a comparison with thick target yields was necessary.
The thick target yields listed in Table 4.2 are portrayed in Fig. 4.2. The proton energy and the
thick target yields are given in the laboratory frame of reference, with energy units in keV and
gamma-ray yield units in number of emitted gamma-rays per sr and per µC.
63
FIG. 4.2. Gamma-ray yield of the 429 keV 7Li(p,nγ1–0)7Be resonance as a function of proton energy,
deduced from the works of a) [4.2] and b) [4.3]. Only statistical errors are shown.
(a) Assessment and recommended data

The measured thick target gamma-ray yields from Kiss et al. [4.2] indicate a proton energy
range comparable to those used in the works from Boni et al. [4.5] and Caciolli et al. [4.6], in
particular the values at 3.1 and 3.8 MeV proton energy. These values were compared with the
corresponding calculated thick target yields obtained from integrating the gamma-ray
producing cross sections measured by Boni et al. and Caciolli et al. between 3.1 and 3.8 MeV
proton energy and making sure to take the proton energy loss into account. As reported in
Table 4.3, it was found that the calculated values from Boni et al. are 12% higher than the
experimental thick-target yield, whereas the calculated values from Caciolli et al. are 18%
lower. The quoted uncertainties were calculated considering the quoted total uncertainties in
the differential cross-section and the uncertainty in the stopping power for protons (4.2%
[4.7]). The calculated yields are in reasonable agreement with the thick target yield, within the
quoted uncertainties, which enabled a recommended cross-section data set for the 429 keV γ-
ray from the 7Li(p,nγ1–0)7Be reaction in the range from 2 to 5 MeV proton energy to be
obtained. This was completed by multiplying the data from Caciolli et al. by 1.18 and the data
from Boni et al. by 0.89. Additional thick target yield measurements should be carried out in
the future to confirm this result.
TABLE 4.3. COMPARISON BETWEEN EXPERIMENTAL THICK TARGET YIELD AND THE
INTEGRATED DIFFERENTIAL CROSS-SECTIONS OF 429 keV 7Li(p,nγ1–0)7Li
Energy interval Thick target yield [4.2] Calculated yield Calculated yield
[4.6] [4.5]
3.8–3.1 MeV 1.68 × 107 (1.42 ± 0.22) × 107 (1.88 ± 0.29) × 107
* γ-ray yields are expressed in units of Nγ/sr/µc
64
4.1.1.2. 7Li(p,pγ1–0)7Li: 478 keV γ-ray
Apart from the data already included in the IBANDL and EXFOR [4.8] compilations, other
papers within the literature present cross-section data in arbitrary or relative units, which
makes them of limited practical use for the PIGE technique. In particular, the paper from Rio
et al. [4.4] presents the emission yields of the 478 keV γ-ray for proton energies from 1 to 3
MeV (Fig. 1c in [4.4]), expressed as counts per µC, whereas the paper from Newson et al.
[4.9] presents the relative yield of the same reaction for energies up to the threshold energy
(Fig. 3 in [4.9]).
(a) New measurements

The Lisbon group measured both the differential cross-section and the thick-target yield of the
7
Li(p,pγ1-0)7Li reaction that emits a 478 keV γ-ray. The experiments were carried out at the
3.0 MV Tandetron accelerator of CTN/IST, using a reaction chamber that is electrically
insulated from the beam line and that works together with the target holder and beam stopper
as one Faraday cup for beam charge collection. The proton beam energy was calibrated with
an accuracy of 0.1 % by using the 1645 keV and 1930.7 keV resonances of the
23
Na(p,p′γ)23Na reaction and the 3470 keV resonance of the 16O(p,p)16O reaction. Gamma-
rays were detected by a HPGe detector with 45% nominal efficiency located at 130° with
respect to the incident beam direction and at a distance of 55.5 mm from the target,
subtending an angle from 100 to 160°.
Measurements of the differential cross-section [4.10] were performed at proton energies from
2.15 to 4.07 MeV, with energy steps ranging from 2 to 20 keV. The target consisted of a thin
film of LiF deposited over a thin self-supporting film of Ag. As the stoichiometry of the target
was impossible to measure, the absolute differential cross-section was obtained by
normalization to previously measured results for energies lower than 2350 keV [4.11]. The
measured cross-section is given as the total cross-section on the basis of the isotropicity of the
478 keV γ-ray emission. The quoted uncertainties are 2% statistical and 7% systematic.
The measurements of the thick target yields [4.10, 4.12] were performed at proton energies in
the range of 1.6 to 4.0 MeV, with proton beam currents to the order of 20 nA. The target was
a pressed pellet made of LiF powder mixed thoroughly with silver powder, yielding an
isotopic 7Li fraction of 3.7%. For this experiment, the absolute efficiency of the HPGe
detector, as a function of the γ-ray energy, was obtained by measuring the yields of γ-rays
produced by radioactive sources (133Ba and 152Eu), calibrated in activity, as well as by Monte
Carlo simulations using GEANT and PENELOPE. The combined uncertainty for the
measurement of the thick target γ-ray yield is 13.5% and can be broken down into the
following components: 1% γ-ray peak area, 6% γ-ray detector absolute efficiency, 10%
absolute value of the collected charge, 5% 7Li fraction and 5% stopping power correction.
(b) Data in the literature and assessment

Table 4.4 lists the cross-section data sets found in the literature and the new cross-section data
measured in the framework of the CRP [4.10]. Table 4.5 reports the thick-target yield data
sets.
65
TABLE 4.4. DATA IN THE LITERATURE FOR THE DIFFERENTIAL CROSS-SECTIONS OF 478 keV 7Li(p,pγ1–0)7Li
Reference Data source Angle Ep (MeV) Target Quoted uncertainties Data presentation
Li evaporated on Be foil
[4.13] EXFOR 90° 0.9–1.4 20% Tabular a
0.2 mg/cm2 metallic Li

[4.14] EXFOR — 1.7–6.0 evaporated on Ta backing 10% Tabular a
1 mg/cm2 metallic Li
[4.15] IBANDL — 3.0–8.0 evaporated on 1 mm Mylar 2%–6% Tabular (R33 file)
backing
1000 Å LiF evaporated on
[4.5] IBANDL 90° 2.2–3.8 Nuclepore filters 15% Tabular (R33 file)
51 µg/cm2 LiF on
[4.11] IBANDL 135° 0.6–2.5 34 µg/cm2 Cu, on 10% Tabular a (R33 file)
15 µg/cm2 C
Thin isotopically enriched
7
[4.16] IBANDL — 1.0–1.8 Li (99.9%) layer 10% Tabular a (R33 file)
evaporated on a Ta disc
50 µg/cm2 LiF on 1–9% statistical,
[4.6] IBANDL 135° 3.0–5.5 30 µg/cm2 C, coated with 13.5% systematic Tabular (R33 file)
20 µg/cm2 Au
Thin LiF film on a thin Ag 2% statistical,
[4.10] IBANDL 130° 2.1–4.1 film 7% systematic Tabular a (R33 file)
a
cross-section values are given as total (in mbarn)
66
TABLE 4.5. DATA IN THE LITERATURE FOR THE THICK TARGET YIELD OF 478 keV 7Li(p,pγ1–0)7Li
Reference Data source Angle Ep (MeV) Target Quoted uncertainties Data presentation
[4.1] IBANDL 55° 1.0–2.4 Li powdered compound — Tabular (R33 file)
pressed into pill,
1 mm thick
[4.2] IBANDL 55° 2.4–4.2 Pellet of Li chemical — Tabular (R33 file)
compound
[4.3] IBANDL 55° 7.0–9.0 Pellet of Li powdered 2–5% statistical, Tabular (R33 file)
chemical compound, 7% systematic
1 mm thick
[4.17] IBANDL 90° 1.0–4.1 Pellets of Li compound — Tabular (R33 file)
(< 10%) mixed with
powder graphite and
cellulose
[4.10, 4.12] IBANDL 130° 1.6–4.0 Pellet of LiF mixed with 1% statistical, Tabular
Ag 13.5% systematic (R33 file)
67
Figure 4.3 illustrates the cross-sections that are listed in Table 4.4. The proton energy and the
keV and cross-section units in mbarn/sr. The values from [4.10], [4.11], [4.13], [4.14] and
[4.16] are given in IBANDL as total cross-section: for the comparison shown in Fig. 4.3, the
isotropic angular distribution of the emitted γ-ray has been taken into account by scaling the
cross-sections down by 4π.
FIG. 4.3. Differential gamma-ray production cross-sections of the 478 keV 7Li(p,pγ1–0)7Li resonance
as a function of proton energy deduced from the works of a)[4.6], b) [4.5], c) [4.10], d) [4.11],
e) [4.16], f [4.15], g [4.13] and h [4.14]. Only statistical errors are shown.
In general there is an overall agreement in shapes between the different data sets. In the
proton energy region below 2 MeV, data from Brown et al. [4.13] are two times higher than
those from Mateus et al. [4.11], whereas data from Aslam et al. [4.16] are about 17–18%
higher than those from Mateus et al. Both the quoted uncertainties are 10%, but appear shifted
in energy around 30 keV.
In the high proton energy region, above 2 MeV, data from Fonseca et al. [4.10] are in obvious
agreement with those from Mateus et al., having been normalized to the latter. Cross-section
values from Boni et al. [4.5] are about 12% higher than those from Fonseca et al. while those
from Caciolli et al. [4.6] are about 10–15% lower. Data from Presser and Bass [4.14] are
about 5–10% higher than data from Fonseca et al. It should be noted that the bump in the
cross-section around 2.9 MeV proton energy and evident in the data from Presser and Bassi,
is not visible in the other data, except, partially, in those from Boni et al. Although scarce, the
data from Guzhovskij et al. [4.15] are in reasonable agreement with those from Presser and
Bass. Considering that the quoted uncertianties for the different cross-section data sets are
around 10–15%, the observed systematic discrepancies between the different data are within
the errors.
68
Figure 4.4 illustrates the thick target yields listed in Table 4.5. The proton energy and the
thick target yields are given in the laboratory frame of reference, with energy units in keV and
FIG. 4.4. Gamma-ray yield of the 478 keV 7Li(p,pγ1–0)7Li resonance as a function of proton energy,
deduced from the works of a) [4.2], b) [4.3], c) [4.1], d) [4.17] and e) [4.10, 4.12]. Only statistical
errors are shown.
Note that the thick target γ-ray yields from Kiss et al. [4.2] are normalized to the yields from
Antilla et al. [4.1] at 2.4 MeV proton energy. Gamma-ray yields from Savidou et al. [4.17]
are 20% to 50% lower than those from Kiss et al. and Antilla et al., with the larger differences
appearing in the low proton energy region (below 2.4 MeV). Gamma-ray yields from Jesus et
al. [4.10, 4.12] are sligthly lower (from 5% to 15%) than those from Savidou et al., but still in
agreement within the quoted uncertainties in the high proton energy range (above 3.6 MeV).
However, the difference increases 15% for the data at 2.4 MeV and 50% for the data at 1.6
MeV. These results indicate that the overall shapes of the three thick-target γ-ray yields
obtained from the measurements of Kiss and Antilla, Savidou and Jesus are different.
The measured thick-target γ-ray yields from [4.1], [4.2], [4.10, 4.12] and [4.17], refer to a
proton energy range from 4.2 to 1.0 MeV, which is comparable to the one used in several
works on cross-section measurements. These yields were compared with corresponding
calculated thick-target yields obtained by integrating the γ-ray producing cross-sections
measured by Fonseca et al., Caciolli et al., Presser and Bass, Boni et al., Mateus et al. and
Aslam et al., taking the proton energy loss into account and using the following proton energy
integration intervals: 4.2–3.8 MeV, 3.8–3.1 MeV, 3.1–2.4 MeV, 2.4–1.7 MeV and 1.7–1.0
MeV. Given the measured data points, the thick-target yields from Jesus et al. will be
considered only for comparison within the integration interval 2.4–1.7 MeV. Gamma-ray
yields from Räisänen et al. [4.3] refer to proton energies that are higher than those used for
measuring the differential gamma-ray production cross-sections of the 478 keV γ-ray.
Therefore they could not be used for the validation.
69
As shown in Table 4.6, it was found that, for proton energies above 3 MeV, the calculated
yield values from Fonseca et al. [4.10] and from Presser and Bass [4.14] are in good
agreement with the measured thick target yields and well within the quoted uncertainties. The
uncertainties are calculated considering the quoted total uncertainties in the differential cross-
section and the uncertainty in the stopping power for protons (4.2% [4.7]). Calculated values
based on the values from Boni et al. [4.5] are higher than measured γ-yields and values from
Caciolli et al. [4.6] are lower than the measured γ-yields.
In the energy interval between 2.4 and 1.7 MeV, the calculated γ-ray yield from the only
existing data set, i.e. Presser and Bass, is in agreement with the experimental thick target
yields from Savidou et al. [4.17], Jesus et al. [4.10, 4.12] and Antilla et al. [4.1], within the
quoted uncertainties. It must be noted that, given the measured data points, the thick-target
yield from the data of Jesus et al. refers to a larger energy range (1.6–2.4 MeV) and therefore
overestimates the yield. However, the difference is less than 10%. Below 1.7 MeV there is a
factor of 2 between the thick target yields from Savidou et al. and Antilla et al., while the
calculated yields using the cross-section values of Mateus et al. [4.11] and Aslam et al. [4.16]
lie just in between.
It is recommended that additional thick target yield measurements be carried out, in particular
for proton energies below 3 MeV, to solve the aforementioned discrepancies.
The data included in this detailed analysis had a cut-off date of December 2016. Since then,
there have been additional publications of new data from CMAM [4.10]. These data were not
included in our assessment; however they are available in IBANDL.
70
TABLE 4.6. COMPARISON BETWEEN EXPERIMENTAL THICK-TARGET YIELDS AND THE INTEGRATED DIFFERENTIAL CROSS-SECTIONS
OF 478 keV 7Li(p,pγ1–0)7Li
Energy Thick Thick Thick Thick Calculated Calculated Calculated Calculated Calculated Calculated
interval target yield target yield target yield target yield yield yield yield yield yield yield
(MeV) [4.2] [4.17] [4.10, 4.12] [4.1] [4.10] [4.6] [4.14] [4.5] [4.11] [4.16]
4.2–3.8 2.10 × 107 — — (1.97 ± (1.99 ± (2.34 ± — — —
0.16) × 107 0.28) × 107 0.25) × 107
3.8–3.1 3.30 × 107 3.25 × 107 — (3.34 ± (2.80 ± (3.40 ± (3.64 ± — —
0.27) × 107 0.23) × 107 0.37) × 107 0.57) × 107
3.1–2.4 3.00 × 107 2.43 × 107 — (2.39 ± — (2.82 ± — — —
0.19) × 107 0.30) × 107
2.4–1.7 — 1.45 × 107 (1.71 ± 1.74 × 107 — — (1.58 ± — — —
0.23) × 107 0.17) × 107
1.7–1.0 — 3.93 × 106 7.95 × 106 — — — — (5.46 ± (6.51 ±
0.59) × 106 0.70) × 106
* γ-ray yields are expressed in units of Nγ/sr/µC
71
4.1.2. Gamma-ray production cross-sections of lithium from deuteron-induced nuclear
reactions
Thick target measurements [4.18] determined that, in the case of lithium, only two γ-rays
from the 6Li isotope are excited by deuterons with energy less than 2 MeV. These are the 429
keV γ-ray from the 6Li(d,nγ1–0)7Be reaction and the 478 keV γ-ray from the 6Li(d,pγ1–0)7Li
reaction. Both γ-rays have exactly isotropic angular distribution. There are no γ-rays
originating from the more abundant 7Li isotope; the 7Li(d,pγ1–0)8Li reaction, which emits a γ-
ray of 981 keV has a negative Q-value of -192 keV and the threshold for the transition from
the first excited level is 1.51 MeV, which means that the yield of the γ-ray is very low for
deuterons with energy less than 2 MeV.
4.1.2.1. 6Li(d,nγ1–0)7Be: 429 keV γ-ray
Table 4.7 lists the only thick-target yield data set that was found for this reaction.
6
Li(d,nγ1–0)7Be
Reference Data Angle Ed (MeV) Target Quoted Data
[4.18] IBANDL 135° 0.7–1.9 LiSO4 thick 7–15% Tabular (R33
sample statistical file)
5%
systematical
Figure 4.5 illustrates the thick target yield listed in Table 4.7. The deuteron energy and the
thick-target yield are given in the laboratory frame of reference, with energy units in keV and
FIG. 4.5. Gamma-ray yield of the 429 keV 6Li(d,nγ1–0)7Be resonance as a function of deuteron energy
deduced from the work of a) [4.18]. Only statistical errors are shown.
72
4.1.2.2. 6Li(d,pγ1–0)7Li: 478 keV γ-ray
After an investigation of the literature, it was found that the only cross-section data was from
Sziki et al. [4.19] and had already been uploaded to IBANDL. Another data set, from
Guzhovskij et al. [4.15], is available in the EXFOR database [4.8]. However, it refers to high
deuteron energies from 3 to 11 MeV, which means that these cross-sections are of limited
practical use for the PIGE technique.
Table 4.8 reports the cross-section data found in the literature and Table 4.9 reports the thick-
target yield data found in the literature.
TABLE 4.8. DATA IN THE LITERATURE FOR THE DIFFERENTIAL CROSS-SECTIONS OF

478 keV 6Li(d,pγ1–0)7Li
[4.19] IBANDL 60° 0.8–2.0 LiF and LiBO2 1–7% Tabular (R33
layer onto Ta statistical file)
substrate 5%
systematical
6
Li(d,pγ1–0)7Li
[4.18] IBANDL 135° 0.8–2.0 LiSO4 thick 2–8% Tabular (R33
sample statistical file)
5%
systematical
Figure 4.6 presents the cross-section listed in Table 4.8. The deuteron energy and the
keV and cross-section units in mb/sr.
73
FIG. 4.6. Differential γ-ray production cross-sections of the 478 keV 6Li(d,pγ1–0)7Li resonance as a
function of deuteron energy deduced from the works of a) [4.19]. Only statistical errors are shown.
Figure 4.7 presents the thick-target yield listed in Table 4.9. The deuteron energy and the
thick-target yield are given in the laboratory frame of reference, with energy units in keV and
γ-ray yield units in number of emitted γ-rays per sr and per µC.
FIG. 4.7. Gamma-ray thick-target yield of the 478 keV 6Li(d,pγ1–0)7Li resonance as a function of
deuteron energy deduced from the works of a) [4.18]. Only statistical errors are shown.
74
(c) Assessment and recommended data
Since the measured thick target γ-ray yields from Elekes et al. [4.18] refer to the same
deuteron energy range used in the work from Sziki et al. [4.19], they were compared with
corresponding calculated thick-target yields obtained by integrating from 1.0 to 2.0 MeV
deuteron energy over the γ-ray producing cross-sections measured by Sziki et al., taking the
proton energy loss into account. As shown in Table 4.10, it was found that the calculated
values from Sziki et al. are in good agreement with the thick-target yield from Elekes et al.,
within the quoted uncertainties. For the calculated yield, the quoted uncertainties were
obtained by considering the quoted total uncertainties in the differential cross-section and the
uncertainty in the stopping power for deuterons (assumed conservatively to be 5%). Based on
these results, the cross-section data set from Sziki et al. for the 478 keV γ-ray from the
6
Li(d,pγ1–0)7Li reaction in the range from 1 to 2 MeV deuteron energy can be recommended.
TABLE 4.10. COMPARISON BETWEEN EXPERIMENTAL THICK-TARGET YIELD AND THE

INTEGRATED DIFFERENTIAL CROSS-SECTIONS OF 478 keV 6Li(d,pγ1–0)7Li
Energy interval Thick target yield Calculated yield
[4.18] [4.19]
2.0–1.0 MeV (3.40 ± 0.25) × 107 (3.02 ± 0.21) × 107
* γ-ray yields are expressed in units of Nγ/sr/mC.
4.2. BERYLLIUM
Beryllium can be detected in PIGE measurements of reactions induced by proton, deuteron

and alpha particles. The γ-ray emission by heavy ion beams such as carbon, nitrogen or
oxygen at energies between 22 and 33 MeV have been reported by Seppälä et al. [4.20], but
are at present beyond the scope of this report and the IBANDL data base. The deuteron and
alpha induced reactions are favorable, because they have a higher sensitivity than the proton-
induced reactions, however at the cost of a neutron background. The proton and deuteron
induced reactions are often used because they enable the simultaneous determination of other
light elements. It should be noted that particle spectroscopy is sometimes favored over the
PIGE technique for the detection of Be because the detectors are less sensitive to neutrons and
some depth information can be obtained.
4.2.1. Gamma-ray production cross-sections of beryllium from proton-induced nuclear

reactions
4.2.1.1. 9Be(p,γ)10B
The 9Be(p,γ)10B reaction has a limited sensitivity for the detection of beryllium [4.21] due to a
low cross-section. The cross-section has been studied for astrophysical purposes in the energy
range of 70 keV to 1.8 MeV by Zahnow et al. [4.22] and is displayed in Fig. 4.8 together with
earlier data [4.23].
75
1.E-03
1.E-04
o a)
cross section [mbarn] x b)
1.E-05
1.E-06
1.E-07
1.E-08
1.E-09
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8
beam energy [MeV]
FIG. 4.8. Cross-section of the 9Be(p,γ)10B reaction as a function of beam energy, deduced from the
works of a) [4.22] and b) [4.23]. Data of a narrow resonance at Ep = 1.087 have been omitted.
The data from Zahnow et al. have been measured with a 4π NaI(Tl) summing crystal that
analyzes the summing peak with energies between 6.5 and 8.3 MeV. The authors quote an
error for the absolute cross-section of 7%, which is caused mainly by the uncertainty of the
detector efficiency.
During the course of the CRP, thick-target yields were measured [4.24] in the beam energy
range between 200 keV and 1.6 MeV, to analyze the 718 keV γ-line from the first excited
state in 10B. These measurements were compared to thick-target yields calculated with the
ERYA code (see Section 5) by an integration of the data from Zahnow et al. (shown in Fig.
4.8). The authors also measured cross-sections for the 718 keV transition at selected energies
and renormalized the excitation curve of Zahnow et al. to these results. The agreement
between the calculated values and the measured thick-target yields is better than 10%,
indicating that the relative cross-section data in Fig. 4.8 are correct within this error.
4.2.1.2. 9Be(p,αγ)6Li
The γ-yield of thick targets has been measured by Meyerhof et al. [4.25] and later, in studies
dedicated to material analysis, by Anttila et al. [4.1] and Kiss et al. [4.2] for different γ-lines
at several beam energies. The results of the latter two are summarized in Table 4.11. (Table 1
in [4.2]). The 3562 keV line from the 9Be(p,αγ)6Li reaction becomes the dominant
contribution for increasing beam energies, leading to better sensitivity. This has been used,
for example, to measure beryllium concentrations in PIXE/PIGE studies of emeralds with an
external proton beam of 3.0 MeV [4.26] and 2.4 MeV [4.27].
76
TABLE 4.11. THICK TARGET YIELDS FROM 3562 keV γ-RAY OF 9Be(p,αγ)6Li AND
9
Be(p,γ)10B (ALL OTHER RESONANCES).
Yield [Nγ/(µC ⋅ sr)]
Eγ (keV)
[4.1] [4.2]
1.0 MeV 1.7 MeV 2.4 MeV 3.1 MeV 3.8 MeV 4.2 MeV
415 0.4 ×103 0.6 × 103 0.6 × 103 — — —
718 1.2 × 103 3.5 × 103 5.3 × 103 — — —
1022 0.4 × 103 0.8 × 103 1.3 × 103 — — —
3562 — 0.1 × 103 2.5 × 104 2.5 × 106 5.1 × 106 6.2 × 106
7477 3.3 × 103 3.0 × 103 — — — —
The data from Table 4.11 have been measured at a detection angle of 55°. No detailed error
discussion is given but the authors estimate that the main error is probably due to the
uncertainties in the stopping power.
4.2.2. Gamma-ray production cross-sections of beryllium from deuteron-induced

nuclear reactions
4.2.2.1. 9Be(d,nγ)10B
Among the deuteron induced reactions, the 9Be(d,nγ)10B reaction is the only γ-ray emitting
one with potential for material analysis. It has the drawback of a relatively high neutron flux,
which is disadvantageous for Germanium detectors and can cause serious deterioration of the
detector properties. Nevertheless it has a high sensitivity for the beryllium detection
comparable to the α-induced reaction (see below).
The first excited state in 10B has spin value 1+, and therefore the emitted γ-ray is expected to
have anisotropic angular distribution. The cross-section for this γ-line 718 keV has been
measured with thin targets in the energy range 600–2 MeV under a detection angle of 60° by
Sziki et al. [4.19]. The data are presented in Fig. 4.9. The error bars in the figure represent the
error in the number of target atoms (< 1%), the charge accumulation (< 0.3 %), the detector
efficiency (3 %) and the stopping power (3%).
77
20
18
16
cross section [mb/sr]
14
12
10
6
0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1
beam energy [MeV]
FIG. 4.9. Cross-section of the 718 keV 9Be(d,nγ)10B resonance at 60° as a function of beam energy in
MeV, deduced from the work of [4.19].
Perdikakis et al. [4.21] have measured thick target yields for beam energies between 600 keV
and 2.1 MeV and also for the 718 keV line, but with a detection angle of 90°. For the beam
energy of 1.5 MeV, they report 1.1 × 107 gammas/(µC⋅sr). The energy dependence is
displayed in Fig. 4.10.
2.5E+07
absolute yield [N/(µCsr)]
2.0E+07
1.5E+07
1.0E+07
5.0E+06
0.0E+00
0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3
beam energy [MeV]
FIG. 4.10. Thick target absolute yield for the 718 keV 9Be(d,nγ)10B resonance at 90° as a function of
beam energy in MeV, as deduced from the work of [4.21].
78
The angular distribution of the 718 keV line was investigated at the beam energy of 1.85 MeV
at angles of 90°, 75° and 60° and found to be isotropic within 15%. In this pioneering work,
the reaction was used to analyze Be in soil, asphalt and concrete, with sensitivities below 100
ppm.
4.2.3. Gamma-ray production cross-sections of beryllium from alpha-induced nuclear

reactions
4.2.3.1. 9Be(a,nγ)12C
The gamma emission of the α-induced reaction on beryllium is for analysis purposes similar
to the case of the deuteron-induced reaction in that it is also hampered by the high neutron
flux and therefore not often used. The deuteron-induced reaction, however, has the advantage
of being sensitive to other light elements. Nevertheless, the alpha-induced reaction has some
importance for the analysis of fusion devices [4.28] and can be used for analysis with
accelerator beams as well. Borderie et al [4.29] explored the analytical potential by estimating
the sensitivity for Be in Nb with a beam energy of 3.5 MeV, but did not report cross-sections
or thick-target yields. The sensitivity was estimated to be 0.5 ppm. Later, Lappaleinen et al.
[4.30] measured the thick-target yield at 2.4 MeV to be 1.6 × 106 gammas/(µC⋅sr) at a
detection angle of 55° and deduced a detection limit in volcanic stone of 0.45 ppm. Future
measurements at various beam energies are recommended.
It should be noted that thick-target yields for beam energies between 5.6 and 10 MeV have
been measured by Heaton et al. [4.31]. These measurements are needed for the calculation of
the background of ultralow-background experiments, but such high beam energies may not be
useful for material analysis.
4.3. BORON
4.3.1. Gamma-ray production cross-sections of boron from proton-induced nuclear

reactions
Boron is the fifth lightest element, consisting of two isotopes: 11B (80.1%) and 10B (19.9%).
Its low Z makes detection almost impossible with the PIXE technique and it can only be
quantified by other IBA techniques (EBS, NRA, etc.) when present in suitable matrices. PIGE
has been proven to be a reliable and widely used method for its detection and quantification.
The most commonly used reactions in PIGE are 10B(p,αγ1–0)7Be, 10B(p,p′γ1–0)10Β and 11B
(p,p′γ1–0)11B, which emit the 429, 718 and 2125 keV gamma-rays, respectively. Despite the
fact that they have been used for over two decades [4.5, 4.32, 4.33], the available cross-
section data contains large discrepencies, with differences that in some cases reach up to a
factor of 5. In the framework of the present CRP, a new set of measurements was considered
necessary in order to untangle these differences.
10
4.3.1.1. B(p,αγ1–0)7Be: 429 keV γ- ray

The 429 keV γ-ray is emitted from the de-excitation of the first excited level of 7Be,
populated through the 10B(p,αγ)7Be reaction. The angular distribution of this γ-ray is expected
to be isotropic since the spin of the first excited state of 7Be is -1/2. A recent measurement of
the differential cross-section of the 10B(p,αγ1–0)7Be reaction was performed by Lagoyannis et
79
al. [4.34] for eight angles and for the proton energy range between 2.0 to 5.0 MeV with a
variable step of 20 to 40 keV. The measurement was performed at the 5.5 MV Tandem
Accelerator installed at the Institute of Nuclear and Particle Physics, NCSR “Demokritos”.
The energy calibration of the accelerator was performed using the narrow resonances of
27
Al(p,γ)28Si and 13C(p,γ)14N at Ep = 991.9 and 1746.9 keV, respectively. Two thin-targets
were used throughout the experiment: an enriched 10B and a natural 10B. Both targets were
produced by electron gun evaporation onto tantalum backings and were characterized by
simultaneously implementing the EBS and NRA techniques. The targets were placed at the
center of a cylindrical chamber, which also acted as a Faraday cup. The beam charge was
collected using a current integrator while a suppression voltage of + 300 volts was applied
onto a tantalum collimator placed ~1 meter before the chamber. The produced γ-rays were
detected using four HPGe detectors placed at initial angles of 0ο, 55ο, 90ο and 165ο with
respect to the beam direction, and at a distance of ~ 25 cm from the target. After each energy
step, the detectors were rotated by 15o to a set of four new angles, i.e. 15o, 40o, 105o and 150o.
The energy and efficiency calibration of the HPGe detectors were performed using a
calibrated point source of 152Eu [4.34].
The statistical error derived from the integration of the accumulated spectrum peaks did not
exceed 5%, but the overall systematic error, which takes into account the uncertainties from
(i) beam charge collection, (ii) efficiency calibration and (iii) target thickness was estimated at
8%. The uncertainty budget can be found in Table 4.12.
TABLE 4.12. UNCERTAINTY BUDGET FOR THE DIFFERENTIAL CROSS- SECTIONS OF 429
keV γ-RAY OF 10B(p,αγ)7Be [4.34]
Peak Integration 3–5%
Beam charge 3.5 %
Target thickness 7%
Detector’s efficiency 3%
The measured differential cross-sections with respect to the proton beam energy are presented
in Fig. 4.11.
FIG. 4.11. Differential cross-sections of the 429 keV 10B(p,αγ)7Be resonance as a function of proton
energy, deduced from the work of [4.34]. Gamma-rays detected from eight different angles with
respect to the beam direction.
80
In addition to the differential cross-section measurements, new thick-target yields were
measured by Chiari et al. [4.35]. Three HPGe detectors were placed at 0o, 45o and 90o with
respect to the beam axis and were calibrated using two point sources; 152Eu and 226Ra. The
proton beam ranged from 2.5 to 4.1 MeV in steps of 150 keV. The accelerator was calibrated
using the resonances at 991.86 keV, 1683.57 and 3470 keV from the 27Al(p,γ)28Si,
27
Al(p,p′γ)27 Al and 16O(p,p)16O reactions, respectively. A voltage of + 300 V was applied to
the target in order to suppress secondary electrons emission. The uncertainty budget of the
experiment is summarized in Table 4.13 and the results are shown in Fig. 4.12.
TABLE 4.13. UNCERTAINTY BUDGET FOR THE THICK-TARGET YIELDS FOR 429 keV γ-
RAY OF 10B(p,αγ)7Be [4.35]
Peak Area 1–2%
Beam charge 4%
Sample Composition 2%
Stopping power 4%
Detector’s absolute efficiency 6%
FIG. 4.12. Thick target yields of the 429 keV 10B(p,αγ)7Be resonance as a function of proton energy,
deduced from the work of [4.35]. Gamma-rays detected at angles of 0o, 45o and 90o with respect to the
beam axis.

In Fig. 4.13, a compilation of the existing cross-section data are presented along with the new
data from Lagoyannis et al. [4.34] at 55o.
81
FIG. 4.13. Comparison of the available differential cross-sections of the 429 keV 10B(p,αγ)7Be
resonance as a function of proton energy. Data from the works of (a) [4.34] at 55°, (b) [4.32],
(c) [4.33] at 90°, (d) [4.36] at 0°, (e) [4.5] at 90°, and (f) [4.37] at 55°.
The data reveals huge discrepancies between the existing datasets and confirms the need for
new measurements, both of which issues were addressed within the framework of this CRP.
The differential cross-sections of the 10B(p,αγ1–0)7Be reaction were measured by Ophel et al.
[4.37] using a 7.6 cm × 7.6 cm sodium iodide crystal placed at 55o for the proton energy range
between 2.4 MeV and 7.2 MeV. The target was composed of a layer of enriched 99.8% 10B
with a thickness of 675 µg/cm2 supported by a tantalum backing. The old data are
approximately 3.5 times lower than the new measurements, a factor which cannot be
explained by the 20% uncertainty given by the authors. However, the authors state that in the
derivation of the cross-section data, they used the nominal thickness of the target as was
provided by the manufacturer and that no additional measurement was performed.
Comparison of the data also reveals a slight energy shift of ~70 keV above Ep = 4.2 MeV.
The measurements from Boni et al. [4.5] are lower by a factor of ~5 compared with the new
data, but there is a high level of agreement with respect to the excitation function. The
detection setup consisted of a HPGe detector placed at 90o and the proton energy range
covered was between 2.2 and 3.8 MeV with a 20 keV step. The target used was composed of
a thin (600 A) layer of natural boron evaporated onto a Nulcepore filter. The systematic
uncertainty, as stated by the authors, is 15%, which cannot explain the existing differences.
Segel et al. [4.32] used a thin enriched 96% metallic 10B target, which was evaporated onto a
carbon foil. The produced gamma-rays were detected by two large NaI detectors placed
perpendicularly to the beam axis and were operated together and separately. The authors
report a rather high dead time (50%) during the measurement: a fact which could explain the
disagreement between the reported cross-sections and those of Lagoyannis et al. [4.34].
There is no clear reason for the disagreement between the new data and those of Hunt et al.
[4.36] since an uncertainty budget in not provided by the authors. For the experiment, they
82
used a 4’’ × 4’’ NaI(Th) detector placed at 0o and a set of thin (10–100 µg/cm2) enriched
(99%) 10B targets evaporated onto copper or tantalum backings.
Finally, there is very good agreement between the new data and the data from Day et al.
[4.33], for the proton energy range between 2.0 and 2.5 MeV. Above 2.5 MeV, Day et al.
claim that there is an overestimation of the reported cross-sections coming from the
overlapping of the 429 keV and 718 keV peaks due to the poor resolution of the NaI detector
used.
Further assessment of the cross-section data was done by comparing the integrated cross-
section data with previously reported thick-target yields in Fig. 4.14. Anttila et al. [4.1]
measured thick-target yields in two proton energies: 1.7 and 2.4 MeV. Additional data for
higher proton energies are provided by Kiss et al. [4.2]. However, the data of Kiss are not
absolute measurements but are normalized to those of Anttila et al. at 2.4 MeV. An
independent measurement by Savidou et al. [4.17] was also included in the comparison, as
was the data of Chiari et al. [4.35]. Since Chiari et al. measured isotopic thick-target yields,
the data had to be normalized to the elemental in order for accurate comparisons to be made
with the data from Antilla, Savidou and Kiss. The integrated cross-section data were
normalized to the closest thick-target yield, with the exception of the data from Lagoyannis et
al. [4.34], for which a normalization factor was not used. The comparison indicates that there
is close agreement between most of the thick-target data and the data of Day et al. and
Lagoyannis et al.
FIG. 4.14. Comparison of the thick-target yields of the 429 keV 10B(p,αγ)7Be resonance as a function
of proton energy. Gamma-rays were detected at 0o and were normalized to elemental yields. Deduced
from the works of (a) [4.34], (b) [4.5], (c) [4.33], (d) [4.37], (e) [4.36], (f) [4.32], (g) [4.17],
(h) [4.1], (i) [4.2] and (j) [4.35].
83
(c) Recommended data
For the proton energy range between 0.94 and 2.0 MeV, the dataset of Day et al. [4.33] is
recommended, since it is in close agreement with the thick-target yield measurements.
Moreover, Mateus et al. [4.38] have applied these data and successfully reproduced the
concentration of boron in various thick samples. For the energy range above 2.0 MeV, the
data of Lagoyannis et al. [4.34] are recommended, since they agree with those of Day et al. in
the common energy range and reproduce the measured thick-target yields by Chiari [4.35],
Kiss [4.2] and Savidou [4.17].
10
4.3.1.2. B(p,pγ1–0)10B: 718 keV γ- ray

The 718 keV γ-ray is the result of the de-excitation from the first excited state of 10B, which is
populated through the 10B(p,p′γ)10B reaction. The angular distribution of this γ-line is
expected to be anisotropic due to the spin 1+ value of the 10B first excited state. While a
number of previous datasets exist in the literature [4.32, 4.33, 4.36, 4.37], there are
disagreements between them, just like there are for the 429 keV γ-ray. A new measurement
was performed by Lagoyannis et al. [4.34] in order to sort out these differences. The
experimental setup used was the same as in the case of the 10B(p,αγ)7Be reaction and all the
experimental details can be found in the section labeled 10B(p,αγ)7Be: 429 keV γ-ray. The
differential cross-sections for all eight angles are presented in Fig. 4.15.
84
FIG. 4.15. Differential cross-sections of the 718 keV 10B(p,p′γ)10B resonance as a function of proton
energy, deduced from the work of [4.34]. Measurements taken at eight different angles with respect to
the beam axis, as noted in the legend within each panel.
As in the case of the 429 keV γ-ray, Chiari et al. [4.35] measured thick-target yields for three
angles. The experimental setup is the same as that described in the section labeled
10
B(p,αγ)7Be: 429 keV γ-ray and results are presented in Fig. 4.16.
85
FIG. 4.16. Thick target yields of the 718 keV 10B(p,p′γ)10B resonance as a function of proton energy,
deduced from the work of [4.35]. Gamma-ray yield was detected at 0o, 45 o and 90o with respect to the
beam axis.

Thick-target yields for the 10B(p,p′γ)10B reaction were measured by Kiss et al. [4.2], Anttila et
al. [4.1] and Savidou et al. [4.17] for the same energies as in the case of the 10B(p,αγ)7Be
reaction. Differential cross-sections were measured by Segel et al. [4.32], Day et al. [4.33],
Hunt et al. [4.36] and Ophel et al. [4.37]. While the new measurements are in excellent
agreement with the data from Day et al., there are major disagreements in regards to the data
from other laboratories. The discrepencies between these datasets can be explained using the
same reasoning as in the case of the 429 keV gamma-rays, in the section labeled
10
B(p,αγ)7Be: 429 keV γ-ray. The datasets are illustrated in Fig. 4.17 for easy visual
comparison.
86
FIG. 4.17. Comparison of the available differential cross-sections of the 718 keV 10B(p,p′γ)10B
resonance as a function of proton energy. Data from: (a) [4.34] at 550, (b) [4.33] at 90°, (c) [4.32],
(d) [4.36] at 0° and (e) [4.37] at 55°.

Since the new measurements of the 718 keV γ-ray are correlated with the 429 keV γ-ray, the
new measurements that were made in the proton energy range from 2.0 to 5.0 MeV are
recommended. For proton energies bellow 2.0 MeV, the data of Day et al. [4.33] are
recommended. The validity of these two datasets is additionally supported by comparing their
integrated γ-ray yields with the thick-target yields found in the literature [4.1, 4.2, 4.17] and
with the new ones by Chiari et al. [4.35], as shown in Fig. 4.18. Concerning thick-target
yields, the new data from Chiari et al. are recommended becasue they are in agreement with
the data from both Day et al. and Lagoyannis et al. [4.34].
87
FIG. 4.18. Comparison of the thick target yields of the 718 keV 10B(p,p′γ)10B resonance as a function
of proton energy. Data obtained from (a) [4.34], (b) [4.33], (c) [4.36], (d) [4.36], (e) [4.32], (f) [4.1],
(g) [4.2], (h) [4.17], and (i) [4.35]. Yields were measured at 0o with respect to the beam axis and
normalized to elemental yields.
10
4.3.1.3. B(p,p′γ)10B : 1020, 1430, 2860 and 4440 γ-rays
These γ-rays were studied at proton energies above 4.0 MeV by Segel et al. [4.32] and are
emitted by the de-excitation of the second, third, fourth and seventh excited state of 10B to the
first (see Fig. 4.19). The experimental setup is the same as that described in the section
labeled 10B(p,αγ)7Be: 429 keV γ-ray. Since there are no other data available for these γ-rays,
no evaluation can be performed.
88
FIG. 4.19. Total cross-section measurements of the (a) 1020 keV (b) 1430 keV (c) 2860 keV and
(d) 4440 keV γ-rays produced by the 10B(p,p′γ)10B reaction as a function of proton energy. All data are
taken from [4.32].
11
4.3.1.4. B(p,pγ1–0)11B: 2125 keV γ-ray
At proton energies above 2.5 MeV the cross-section of the 11B(p,p′γ1–0)11B becomes important
and can be used for the PIGE technique. The first excited state of 11B (T1/2 = 3.8 fs) emits an
isotropic γ-ray of 2125 keV. As illustrated in Fig. 4.20, the 11B(p,pγ1–0)11B reaction appears to
have a double peak, which is due to a combination of the Doppler shift and that the reaction
has a double kinematic solution.
89
FIG. 4.20. The shape of the 2125 keV gamma-ray at different proton energies as noted in the legend.

The only known differential cross-section measurement was made by Boni et al. [4.5] at an
angle of 90o and for the energy range 2.6 to 3.8 MeV. The lack of experimental data, as well
as the underestimation of the cross-sections for the proton-induced reactions of 10B, which
were measured with the same experimental setup, led to the necessity of a new experiment by
Preketes-Sigalas et al. [4.39]. The experimental setup consisted of three HPGe mounted on a
motorized turntable and initially placed at 0o, 90o and 165o. The proton energy varied from
2.58 to 5.0 keV with an energy step of 20 keV. After each energy step, the turntable was
rotated by 15o, thus placing the detectors in three new angles, i.e. 15o, 105o and 150o. The
accuracy in the determination of the energy was 0.1% since it was measured using the narrow
resonances of 27Al(p,γ)28Si and 13C(p,γ)14N at Ep = 991.9 and 1746.9 keV, respectively. A thin
nat
B target mounted on a Ta backing was used throughout the experiment. The thickness of the
target was found to be 577 × 1015 at/cm2 by using a combination of the RBS and NRA
techniques. The results of this experiment are shown in Fig. 4.21 amd the uncertainty budget
is summarized in Table 4.14.
Since there were very few datasets concerning thick-target yields for the 2125 keV γ-ray,
Chiari et al. [4.35] performed a new experiment using the experimental conditions described
in the section labeled 10B(p,αγ)7Be: 429 keV γ-ray and using a thick BN target. The results
are presented in Fig. 4.22 and the uncertainty budget of this experiment can be found in Table
4.14.
90
FIG. 4.21. Differential cross-sections of the 2125 keV 11B(p,pγ1–0)11B resonance as a function of
proton energy, deduced from [4.39]. Yield measured at 15̊, 105̊ and 150̊ with respect to the beam axis.
TABLE 4.14. UNCERTAINTY BUDGET FOR THE NEW DIFFERENTIAL CROSS-SECTION

MEASUREMENTS OF 11B(p,pγ1–0)11B.
Beam charge 3.5 %
Target thickness 5%
91
FIG. 4.22. Thick-target yields of the 2125 keV 11B(p,pγ1–0)11B resonance as a function of proton
energy. Data from [4.35]. Yields measured at 0̊, 45̊ and 90̊, with respect to the beam axis.

As previously stated, the only existing data in the literature is from Boni et al. [4.5], measured
at 90o for the energy range between 2.6 and 3.8 MeV. Details of the experimental setup can be
found in section labeled 10B(p,αγ)7Be: 429 keV γ-ray. The differential cross-sections shown
in Fig. 4.23 are a factor of ~5 lower than the new ones. There is no apparent reason for this
difference and the lack of additional data means that there is not enough information to
recommend one set of data over the other. On the other hand, the shape of the integrated
yields obtained from the new measurements are in agreement with the thick-target yields from
Kiss [4.2] and Savidou [4.17], while the ones from Boni et al. exhibit a different trend (see
Fig. 4.24).
92
FIG. 4.23. Comparison of the differential cross-sections of the 2125 keV 11B(p,pγ1–0)11B resonance
between the two available datasets, recorded at 90o with respect to the beam axis, as a function of
proton energy. Data taken from a) [4.39] and b) [4.5].
FIG. 4.24. Comparison of the thick target yields of the 2125 keV 11B(p,pγ1–0)11B resonance as a
function of proton energy. Data taken from (a) [4.39], (b) [4.5], (c) [4.17], (d) [4.2] and (e) [4.35],
measured at 0o and normalized to elemental yields.
93
The data of Preketes – Sigalas et al. are recommended, since they successfully reproduce the
trend of measured thick-target yields from earlier works [4.2], as well as the ones from Chiari
et al. [4.35].
4.3.2. Gamma-ray production cross-sections of boron from deuteron-induced nuclear

reactions
Contrary to the proton-induced reactions with boron, there are only limited cross-section data
for the deuteron-induced reactions. The higher γ-ray yields of the proton-induced reactions
compared to the deuteron ones are the main reason for this lack of data. Only the 11B isotope
has been studied via deuterons and more specifically, the two first excited states of 12B though
the 11B(d,pγ)12B reaction.
11
4.3.2.1. B(d,pγ)12B: 953 and 1674 keV γ- rays
These γ-rays originate from the 11B(d,pγ1–0)12B and 11B(d,pγ2–0)12B transitions, respectively,
and are expected to have anisotropic angular distributions. Their cross-sections at 60o have
been studied by Sziki at al. [4.19] for the deuteron energies between 0.6 and 2.0 MeV. The
energy calibration of the accelerator was performed using the narrow resonance at 991.9 keV
of the 27Al(p,γ)28Si reaction and the γ-rays were detected by a HPGe detector placed at an
angle of 60o and at a distance of 10 cm from the target. Various thin targets were used with
thicknesses around 800×1015 at/cm2. The resulting differential cross-sections along with their
corresponding errors are presented in Fig. 4.25. The thick-target γ-ray yield of the same
reaction was investigated by Elekes et al. [4.18] using a HPGe detector placed at the same
distance from the target, but at an angle of 135o. The thick-target yields are shown in Fig.
4.26. In the absence of any other experimental data,additional measurements are
recommended, especially with thick-targets, for validation purposes.
FIG. 4.25. Differential cross-sections of the (a) 953 and (b) 1674 keV γ-rays produced by the
11
B(d,pγ)12B reaction as a function of proton energy, measured at an angle of 55°. Data from [4.19]
94
11
FIG. 4.26. Thick target yields of the (a) 953 and (b) 1674 keV γ-rays produced by the B(d,pγ)12B
reaction as a function of proton energy, measured at an angle of 135°. Data from [4.18].
4.4. CARBON
Particle-induced γ-ray reactions with carbon isotopes prove most useful for material analysis
purposes when induced with protons or deuterons. When using proton beams with 12C and 13C
in the lower energy region, (p,γ) reactions occur; while at higher energies, the (p,p′γ) reactions
are dominant. If a laboratory can produce a deuteron beam, however, the deuteron-induced
12
C (d,pγ) reaction is more frequently used and is more important. In the following sub-
sections, the methods and parameters for each reaction are discussed, i.e. energy dependence
of the cross-section, typical beam energies used, γ-ray energies and resonances that can be
employed for depth profiling. Examples of studies within material science are provided in
order to illustrate the use and capabilities of each specific reaction. At the end, the available
data are presented and assessed.
4.4.1. Gamma-ray production cross-sections of carbon from proton-induced nuclear

reactions
4.4.1.1. 12
C(p,γ)13N
This reaction is governed by a direct capture process that interferes with two broad resonances
at 457 keV and 1699 keV for beam energies of up to 2.5 MeV (Fig. 4.27). The primary γ-rays
that are emitted from de-excitation to the ground and first excited state have the energy
provided by the sum of the Q value of the reaction and the proton beam, which is equivalent
to 1.6 MeV and 3.422 MeV, respectively, at a beam energy of 1.056 MeV. Therefore, in
practice, these primary lines are not narrow, but rather broad, depending also on the energy
loss of the protons in the sample. The shape of the line is given by a convolution of the depth
profile of 12C and the cross-section of the reaction. The width of the resonance at 457 keV
allows for a moderate depth resolution, a technique which has been used successfully for the
95
depth profiling of carbon in steel [4.40]. The sensitivity to this reaction is less than the
sensitivity to the competitive (p, p′) reaction at higher energies and to the deuteron-induced
reactions. The advantage, however, is that at low beam energies, background contributions
from other constituents of the sample are small. This method has been used in [4.41, 4.42] for
tracing studies of carbon. The cross-section of the capture reaction to the ground state has
been investigated at the angles of 0° and 90° over a wide range of energies by Rolfs and
Azuma [4.43] in 1974. The results for 90° are displayed in Fig. 4.27.
1.E-02
ccross section [mb/sr]
1.E-03
1.E-04
1.E-05
1.E-06
1.E-07
0 0.5 1 1.5 2 2.5
incident energy [MeV]
FIG. 4.27. Cross-section of 12C(p,γ)13N for a detection angle of 90° as a function of incident beam
energy. Data taken from the work of [4.43].
In the work of [4.43], the absolute cross-section was determined relative to the Rutherford
cross-section using a carbon foil target. With this approach, the measurement is independent
from the target stoichiometry and stopping power data, and only an additional relative
measurement of the beam charge is necessary, which reduces the overall uncertainty. The
majority of the uncertainty comes from the efficiency calibration of the γ-detector. The
authors quote an error of 12 % for the absolute determination of a cross-section of 125 µb at
the maximum of the 457 keV resonance.
More recent measurements at selected beam energies have been performed by [4.44],
resulting in a peak cross-section of 124 ± 12 µb at 460 keV. Since the reaction is important
for nuclear astrophysics, the energy dependence of the cross-section has been investigated
thoroughly by detailed R-matrix calculations [4.44, 4.45]. To improve knowledge of the
reaction rates, a new absolute measurement of the cross-section was initiated during the
course of the CRP, with the goal to further reduce the error. This work is still under way
[4.45].
For measurements on thick samples, the γ-yields for the first (Eγ = 2.366 MeV) and second
(Eγ = 3.511 MeV) excited state of 13N have been measured at beam energies of 1.0 MeV, 1.7
MeV and 2.4 MeV by Anttila et al. [4.1] and at 2.4 MeV and 3.1 MeV by Kiss et al. [4.2].
From the measured spectra provided by [4.1, 4.2], it is clear that these γ-rays are suitable for
96
material analysis. Both measurements were performed at a detection angle of 55 °. The results
are summarized in Table 4.15.
TABLE 4.15. GAMMA-YIELD AT GAMMA ENERGIES 2.366 MeV AND 3.511 MeV FOR
12
C(p,γ)13N AT DIFFERENT BEAM ENERGIES AT AN ANGLE OF 55°.
γ-yield (Nγ/(µC⋅sr))
Eγ (MeV)
1.0 MeV a 1.7 MeV a 2.4 MeV a 2.4 MeV b 3.1 MeV b
2.366 260 220 230 230 270
3.511 — 230 210 210 —
a
[4.1]
b
[4.2]
4.4.1.2. 13
C(p,γ)14N
Due to the low cross-section and the low abundance of 13C found in nature, the 13C(p,γ)14N
reaction is only used in special cases [4.41, 4.42], e.g. for ion beam analysis below beam
energies of 1 MeV. The reaction is governed by a broad resonance at 558 keV and is overall
similar to the 12C(p,γ)13N reaction. Since the Q value is much higher than for the 12C(p,γ)13N
reaction, several primary γ-lines with energies up to more than 8 MeV and several secondary
transitions are available for analysis, which have the advantage of being isotropic when they
come from the first excited state. The cross-section and angular distributions have been
studied in this energy range by King et al. [4.46]. The absolute cross-section was determined
using inverse kinematics relative to a resonance in the 1H (19F,αγ)16O reaction employing a
hydrogen gas target and relative to a resonance in 14N(p,γ) 15O. The authors quote an overall
error of 12%. More recent measurements [4.47] are in good agreement with these results and
give angular distributions at selected proton energies.
Despite being only rarely used, the reaction is important for materials analysis because it has a
strong, narrow and almost completely isolated resonance at a beam energy of 1747 keV,
which is ideally suited for high-resolution depth profiling. Several γ-ray lines are appropriate
for material analysis purposes. Since the γ-line to the ground state has an energy of 9.169
MeV, interferences originating from γ-lines emitted from reactions of other elements possibly
contained in the sample can be avoided. The resonance has been used in various studies such
as isotopic tracing investigations of fusion reactor walls [4.48] and the study of implantation
profiles in silver [4.49] and silicon [4.50].
The resonance parameters, such as branching ratios and angular distributions, are known very
accurately [4.51]. Therefore this resonance is proposed as an alternative to the 992 keV
resonance of 27Al(p,γ)28Si, for the relative efficiency calibration of detectors and the energy
calibration of the beam. Since no recent data on the resonance strength or the cross-section
can be found in literature for this resonance, it is used only in combination with standards.
4.4.1.3. 12
C(p,pγ1–0)12C: 4400 keV γ-ray
The 12C(p,p′γ)12C inelastic scattering has been studied over a wider range of energies, e.g. by
Dyer et al. [4.52], by analyzing the 4.4 MeV γ-ray from the first excited state. The spin value
of the latter is 2+. Therefore, the angular distribution of the γ-ray is expected to be
anisotropic. The data in Fig. 4.28 show a strong resonance at a proton energy of 5.370 MeV.
In the energy range above this resonance, the sensitivity is high. However, for much higher
97
energies (> 8 MeV) the gamma and neutron background often becomes a severe problem. The
reaction has been employed for the investigation of carbon in steel [4.53, 4.54], where
sensitivities of 10 ppm of carbon have been accomplished. It has also been used for the
determination of carbon in biomedical samples [4.55].
500
450
400
350
cross section [mb]
300
250
200
150
100
50
0
5 6 7 8 9 10 11 12
lab proton energy [MeV]
12
FIG. 4.28. Cross-section of inelastic scattering of protons on a C target as a function of proton
energy. Data from [4.52].
In the measurements of [4.52], the absolute value for the inelastic scattering cross-section was
determined from detailed measurements of angular distributions of the particle exit channel
relative to the Rutherford cross-section, using a thin foil target. This method cancels out
errors in the beam charge integration and the determination of the absolute efficiency of the γ-
detector. An error of 5% for the absolute cross-section was obtained. Unfortunately, detailed
information about the angular distribution measured during the course of the experiment is
not given in [4.52].
4.4.1.4. 13
C(p,p′γ)13C: 3090 keV and 3680 keV γ-rays
Inducing the reaction with a low energy beam is not ideal for material analysis, because of the
moderate sensitivity. Two γ-rays emitted from this reaction are normally observed, at 3.09
MeV and 3.68 MeV, with the first one being isotropic. In samples containing traces of
12
C,these γ-rays will be obscured by the strong 4.4 MeV line emitted from the 12C(p,p′γ)12C
reaction at beam energies above 5 MeV. Therefore, the use of these γ-rays is limited to beam
energies below 5 MeV.
98
FIG. 4.29. Relative yield for the 3.09 MeV γ-line (solid points) and the 3.68 MeV line (open points)
produced by 13C(p,p′γ)13C as a function of proton energy. Reproduced courtesy of Physical Review
(1893–1969) [4.56].
The energy dependence of the reaction, as measured by Bair et al. [4.56], is displayed in Fig.
4.29. The authors give relative yields only. Fortunately, Kiss et al. [4.2] have measured the
absolute yield for a thick-target at the beam energies of 3.8 MeV and 4.2 MeV for the 3.09
MeV γ-line. These values were calculated to be 6.2 × 103 and 4.1 × 104 Nγ/(µC⋅sr),
respectively, and can be used to normalize the relative yields of [4.56].
4.4.2. Gamma-ray production cross-sections of carbon from deuteron-induced nuclear

reactions
4.4.2.1. 12
C(d,pγ1–0)13C: 3089 keV γ-ray
The deuteron-induced γ-emission on 12C is generally the preferred method for the detection of
carbon in material analysis because it has a high detection efficiency and the deuteron beam
allows for the simultaneous detection of other light elements such as nitrogen and oxygen.
The technique has many applications, such as for the detection of carbon in steel [4.54] or the
investigation of layers produced by metal-organic chemical vapor deposition (MOCVD)
[4.57].
The dominant γ-ray is emitted from the first excited state in 13C with an energy of 3.089 MeV.
For beam energies above 1.5 MeV, the γ-lines from the second and third state can be seen, but
they are weaker and not used in material analysis. The 3.089 MeV γ-line has the additional
advantage of being isotropic.
99
The differential cross-sections have been studied by several groups, with the most recent
measurements given by Csedreki et al. [4.59], which were performed during this CRP. In this
paper a complete and careful comparison with previous data is presented, an assessment of
the data is given and a benchmarking measurement is performed.
The experiments were performed at ATOMKI using the 5 MV Van de Graaff accelerator. The
energy calibration of the beam was determined with an uncertainty of 3 keV. The
experimental set-up consisted of a target chamber with a long Faraday cup, a co-axial HPGe
detector of 170 cm3 volume positioned at an angle of 55o relative to the beam direction. This
detector was shielded from γ-rays originating from the Faraday cup. The absolute efficiency
was determined with calibrated radioactive sources (133Ba, 56Co, 60Co, 137Cs and 152Eu).
Backscattered particles for an RBS analysis with an alpha beam as well as the particles from
the reaction with a deuteron beam were observed at an angle of 135o with an ion-implanted
silicon particle detector that has a 500 µm active depth and 13 keV energy resolution.
The target was a thin self-supporting natural carbon foil with an evaporated palladium layer
on its back surface. The number of target nuclides was determined with the α-RBS technique.
The simultaneous RBS monitoring of the Pd layer made it possible to determine the absolute
charge independently from a charge measurement from within the target chamber. The γ-ray
yield measurements were performed in the 740–2000 keV deuteron energy range with 2–20
keV steps, depending on the structure of the excitation function. The obtained γ-ray
production cross-sections, as a function of the bombarding deuteron energy are presented in
Fig. 4.30 along with the results of previous measurements.
FIG. 4.30. Differential γ-ray production cross-sections of the 3089 keV γ-ray produced by 12C(d,pγ)13C
as a function of deuteron energy, deduced from the work of a) [4.59] and literature data b) [4.60], c)
[4.61] and d) [4.62].
100
(b) Data from the literature and assessment
There are three previous measurements reported in the literature in which the γ-ray production
cross-sections were measured in a large deuteron energy interval. Both of the studies by Tryti
et al. [4.60, 4.61] primarily explored the nuclear physics aspects of this reaction, while the
aim of the third investigation [4.62] was the application of the results for elemental analysis.
In the measurements of Tryti, the electrostatic beam energy analyzer of the Van de Graaff
generator was calibrated relative to the 13C(p,γ)14N: 1747 keV resonance. The γ-rays were
detected with a planar 10 cm3 Ge(Li) detector. The carbon target was a sooted tantalum
backing. No other details were given for the first measurement [4.60]. In [4.62], the targets
were carbon films of 9 keV thickness for 2 MeV deuterons covered with a thin gold layer.
Particles were detected by two surface barrier detectors at angles of 90° and 135°. The
elastically scattered deuterons from gold were used as a beam current monitor and the target
stability was controlled by comparing the yield from carbon and gold at regular intervals.
Papillon et al. used a 2 MV Pelletron tandem accelerator [4.61]. Its generating voltmeter was
calibrated with the well-known 27Al(p,γ)28Si resonance at 992 keV. The integrated charge was
measured in the insulated chamber with a low-impedance current integrator. For the target, a
thin layer of carbon was deposited onto a solid Ta backing. The detector was a HPGe with an
efficiency of 22.7%. The absolute efficiency of the detector, below 1408 keV, was determined
with a152Eu source placed in the position of the target, while above this energy the efficiency
was obtained from an extrapolation. At 3.1 MeV, the uncertainty of the efficiency was taken
as 5−10%. Table 4.16 shows the uncertainty budget of the new measurement in comparison
with the above two literature data. The final uncertainties for the new measurement are
significantly smaller than the ones from earlier measurements.
TABLE 4.16. COMPARISON OF THE ESTIMATED UNCERTAINTY BUDGET FOR

DIFFERENTIAL CROSS-SECTION VALUES FROM THE MEASUREMENTS OF CSEDREKI et
al. [4.59] AND LITERATURE DATA FOR 3089 keV 12C(d,pγ)13C
Quantity [4.60, 4.61] [4.62] [4.59]
Beam energy spread (keV) ~0.6 1.0
Straggling in the target
Charge collection not given —
Target inhomogeneity 2–5%
NT number of target nuclei per square
5% 3.0%
centimeter
γ-ray peak area 1–2% < 1–2%
γ-ray detector absolute efficiency 5–10% 2.0%
particle detector solid angle — 2.5%
d+Au/Pd elastic peak area — < 1.0%
d+Au/Pd Rutherford cross sect. — 2.0%
Final uncertainty 12% 10% 6%
The experimental conditions of the new [4.59] and three previous works [4.60– 4.62] are
given in Table 4.17.
101
TABLE 4.17. EXPERIMENTAL CONDITIONS OF THE NEW [4.59] AND THREE PREVIOUS
MEASUREMENTS FOR 3089 keV 12C(d,pγ)13C
Parameter [4.60] [4.61] [4.62] [4.59]
Target thickness (at./cm2) ~0.36 х1018* ~2.00х1018* 0.52х1018 1.90 х 1018
Energy loss in target 2.1 11.2 2.9 10.5
at Ed = 1.45 MeV (keV)
Target backing Ta Ta Ta self supp.
Target other — Au — Pd
Target current (nA) 1000 700 70–130 25
Method of charge collection not given Au-RBS Brookhaven IC Pd-RBS
Collected charge (µC) 1800 not given 30 7
Target-detector distance (cm) 15 16 7 9.5
Ge detector volume planar 10 cm3 planar 10 cm3 115 cm3 170 cm3
Ge detector angle 0° 0° 135° 55°
Particle detector no 90° and 135° no 135°
Ed range (MeV) 0.8–2.2 1.4–3.2 0.5–4.0 0.74–2.0
Cross-section unit arbitrary mbarn mbarn mbarn
Uncertainty not given 12% 10% 6%
As it is illustrated in Table 4.17, Tryti et al. [4.61] gave their cross-section values in arbitrary
units in their first paper. Because the experimental conditions in the second measurement
were practically the same as in the first and because there is a large overlap between the
deuteron energy intervals, it was possible to normalize the two curves and then use them for
comparison with other results.
Figure 4.30 shows substantial differences with regard to the energy scale as well as the cross-
section values between the measurements. The energy shift in the data of [4.62] is most likely
caused by an error in the energy calibration of the accelerator. If the data are shifted in
energy, they agree, for beam energies below 2 MeV, with the measurements of [4.59]. For the
energy range above 2 MeV, the situation is not so satisfactory and needs further investigation.
However, work with deuteron beams is mostly performed below 2 MeV, due to an increasing
neutron background at higher energies and for radiation safety reasons.
Thick-target yields for the analysis of infinite thick samples has been measured by Elekes et
al. [4.18] in the beam energy range of 0.7 to 3.4 MeV and an explicit value is given with an
error of ~5 % for the beam energy of 1.8 MeV. These data are in an excellent agreement with
the results from integrating the thin-target differential data from [4.59].

Based on the assessed experimental differential cross-section data, the calculation of an
evaluated cross-section function would be advisable, at least for the 0.7−2.0 MeV deuteron
energy range. For the energy range above 2 MeV, further cross-section measurements are
needed, as mentioned above.
102
4.4.2.2. 13
C(d,pγ)14C
The deuteron induced γ-emission on 13C has a low cross-section and is therefore not of
importance for material analysis. Though the γ-energies of the first and third excited state in
14
C are higher than those from the 12C(d,pγ)13C reaction, they will be practically covered by γ-
lines from other light constituents in the given sample. Very few cross-section data exist in
the literature. Elekes et al. [4.18] gave a γ-spectrum at a deuteron energy of 1.8 MeV on a
solid natural carbon sample, where the lines from the 13C(d,pγ)14C reaction can be seen. No
value for the cross-section, however, is given.
4.5. NITROGEN
4.5.1. Gamma-ray production cross-sections of nitrogen from proton-induced nuclear

reactions
Previous thick-target yield measurements on natural nitrogen showed that there are essentially
two γ-rays which have a yield large enough to be suitable for analytical purposes [4.2]. These
are the 2313 keV γ-ray from the 14N(p,pγ1–0)14N reaction and the 4439 keV γ-ray from the
15
N(p,αγ1–0)12C reaction. The (p,γ) radiative capture reactions have only astrophysical
significance. There are only a few previous thin-target cross-section measurements for
nitrogen, which is likely due to the low yields of the above-mentioned γ-rays. The aim of
these previous studies was more to study basic nuclear physics than to apply them to ion
beam analysis. Therefore, it was important to have new thin-target cross-section
measurements included within the CRP [4.63].
14
4.5.1.1. N(p,pγ1–0)14N: 2313 keV γ-ray

Since the Ex = 2.31 MeV excited state of 14N has a 0 spin value, the angular distribution of
that γ-ray is isotropic, and this excited state has a τm = 98.7 fs mean life causing a Doppler
broadening in the gamma spectrum [4.64]. Recent γ-ray production cross-section
measurements were performed by the Helsinki group and were published in Marchand et al.
[4.63]. The proton beam was generated by the 5 MV belt-driven tandem accelerator of the
University of Helsinki. The 14N(p,p′γ)14N reaction with resonances at 3903, 3996, and
5937 keV were exploited in the calibration of the absolute proton beam energy at 4–6 MeV,
with precision better than 1 keV. The calibration procedure is described in detail in [4.65].
The γ-rays were detected with a 38% efficient HPGe germanium detector positioned 20 mm
from the target. The detection angle was fixed to 55° relative to the beam direction. The
absolute efficiency curve for the detector was determined by using Eu-152, Co-60 and Co-56
calibration sources placed at the position of the target. For more details on energy calibration,
see Section 3.1.
A thin self-supporting Si3N4 membrane, obtained from Silson Ltd, was used as the target for
cross-section measurements. The nominal foil thickness was 100 nm ± 10% and the window
area of the membrane was 5 × 5 mm2. The composition and areal density of the target was
determined accurately by ERDA. The collected charge was measured from a shielded Faraday
cup equipped with secondary-electron suppression. The excitation function was measured in
the proton energy range of 3586–6920 keV with 100 keV energy steps for the regions that are
farther away from the reaction resonance and narrower steps for regions that are closer to the
reaction resonance. The obtained γ-ray production cross-sections are presented in Fig. 4.31 as
103
a function of proton energy. Special care was taken to accurately determine the composition
and areal density of the Si3N4 membrane. Possible sources of error in the measured yield
values include current integration, detection efficiency and the stopping power correction.
FIG. 4.31. Differential γ-ray production cross-sections of the 2313 keV γ-ray 14N(p,pγ1–0)14N
resonance as a function of proton energy deduced from the work of b) [4.63] and from the literature
a) [4.66], c) [4.67] and d) [4.52].

In Fig. 4.31, the results of the new measurement and also the values of three previous cross-
section measurements are shown.
Phillips et al. [4.66] used melanin (C3H6N6) as a nitrogen target, which was evaporated onto
thin carbon backings. No other details about the target were given. The excitation curve was
taken from 3750 to 5220 keV proton energies with steps around 15 keV with a NaI(Tl)
detector. Later, the measurement was repeated but with smaller steps and extended to a proton
energy of 6360 keV using a Ge(Li) detector placed at 90° with respect to the beam direction.
The curve was normalized to an inelastic cross-section of 61 mb for the 3.90 MeV resonance
[4.68]. Systematic uncertainty for the normalization was estimated at less than 10%.
Dyer et al. [4.52] performed γ-ray producing cross-sections measurements that are relevant to
γ-ray astronomy including for the 14N(p,p′γ)14N reaction. Experimental details are as follows.
Proton energies varied from 3.7 to 23 MeV with an energy resolution of a few keV. Charge
determination was performed through the integration of the beam in a Faraday cup. The γ-
rays were detected by two Ge(Li) detectors at ~31o and 110o with respect to the beam
direction. For the target, a gas cell was filled with high purity natural nitrogen. Uncertainties
arise from yield determination (generally 5%), beam current integration (~ 3%), target
104
thickness measurement (5%) and Ge(Li) efficiency measurement (estimated at 5%). These
uncertainties were added quadratically to give the approximate absolute uncertainty of 10%.
Table 4.18 shows the uncertainty budget of the new measurement in comparison with the two
literature data mentioned above. Final uncertainties are in the same range for all three results.
A recent paper, the aim of which was to produce data for nuclear reaction models [4.67], also
contained measured γ-ray production cross-sections of nitrogen in the proton energy range
6550–26200 keV, using a gas target. The γ-rays were detected with four high-purity Ge
detectors between 30◦ and 135◦. Possible sources of error, including from the gas cell and
gamma detector efficiency, were carefully studied. The deduced cross-section values are also
presented in Fig. 4.31 (10 MeV was set as an upper limit of proton energy, which is a usual
limit in IBA applications).
TABLE 4.18. ESTIMATED UNCERTAINTY BUDGET FOR THE ABSOLUTE DIFFERENTIAL

CROSS-SECTION VALUES FOR 2313 keV 14N(p,pγ1-0)14N FROM THE MEASUREMENT OF
[4.63] AND LITERATURE DATA.
Quantity [4.66] [4.52] [4.63]
Beam energy spread (keV) „a few” ~2.5 keV
Charge collection 3% 3%
Target inhomogeneity 5% < 1.0%
NT number of target nuclei per square
— 3%
centimeter
Target stoichiometry — 0.5%
γ-ray peak area 5% 1.5*–3.0**%
γ-ray detector absolute efficiency 5% 5%
Final uncertainty 10% 10% 10%
* at resonance
** off-resonance
The comparison of the measured cross-section data shows that the present values of
Marchand et al. [4.63] are generally higher throughout the entire energy range than the data of
Phillips et al. [4.66] and Dyer et al. [4.52]. However, up to energies of 5 MeV, the
discrepancies do not exceed the experimental uncertainties. Two of the measured cross-
section values of Benhabiles-Mezhoud et al. [4.67] are in the range of Marchand’s values and
match up well. Between these data and the data of Dyer et al., however, there is a large
discrepancy within the energy range 7−8.5 MeV. In the new measurements [4.63], the
resonances are narrower and stronger than in the previous ones.
To check the reliability of thin-target cross-sections, Marchand et al. made thick-target

measurements on different nitrogen containing samples (BN and Si3N4) at several proton
energies. The measured thick-target yields were compared with corresponding calculated
thick-target yields deduced from γ-ray producing cross-sections that were measured by the
same group and Phillips et al. The results are presented in Fig. 4.32. For the sake of
completeness, Fig. 4.32 also shows the results of an old absolute thick-target γ-yield
measurement made by the Helsinki group [4.3] and a very new measurement of Chiari et al.
[4.69]. It was found that the trends for the experimental and calculated values are similar.
However, the present experimental values are systematically lower than the calculated thick-
target yields by 45–30%. The experimental thick-target yields obtained by the Si3N4 target are
105
systematically slightly lower than the values obtained by the BN target. The values obtained
by using the cross-section data of Phillips et al. are in good agreement with the values
obtained by calculating the thick-target yields from the present cross-section data.
FIG. 4.32. Comparison of measured thick-target yields of the 2313 keV γ-ray 14N(p,pγ1–0)14N
resonance as a function of proton energy: (a) BN exp [4.63], b) Si3N4 exp [4.63], with calculated
thick-target yields deduced from the corresponding γ-ray producing cross-sections measured by
c) [4.66] and d) [4.63]. Some measured thick-targets yields from literature are also included for
comparison: e) [4.2], f) [4.3] and g) [4.69].

Based on the results of the above four γ-ray production cross-sections, it can be concluded
that the measurements of Marchand et al. [4.63] are recommended for PIGE analysis due to
the narrower resonances and good agreement with other data, for proton energy up to 6.5
MeV. The large discrepancies between the experimental values above 6.5 MeV suggests that
re-measurement of these cross-sections is advisable.
14
4.5.1.2. N(p,pγ2–1)14N: 1635 keV γ-ray
(a) Data in the literature and assessment

The transition from the second excited state (Ex = 3.95 MeV) to the first one produces the
1635 keV γ-ray. Its’ cross-section as a function of proton energy was extracted from the
measurements of Phillips et al. [4.66], Dyer et al. [4.52] and Benhabiles-Mezhoud et al.
[4.67]. Their results are compared in Fig. 4.33. As demonstrated in Fig. 4.33, there is good
agreement between the values of Phillips et al. and Dyer et al. below 6.5 MeV. The measured
cross-section values of Dyer et al. and Benhabiles-Mezhoud et al. show large discrepancies
above this energy. The 3.95 MeV state has a spin value of 1, which means that the angular
distribution of the 1635 keV γ-ray is anisotropic. Because of the higher reaction threshold, its
106
yield becomes measurable around 5 MeV, and its cross-section is about 50% of the 2313 keV
γ-ray at higher energies.
FIG. 4.33. Differential γ-ray production cross-sections of the 1635 keV 14N(p,pγ2–1)14N resonance as a
function of proton energy, deduced from the work of a) [4.66] at 90°, c) [4.67] and d) [4.52].
(b) Recommended data

The laboratories that have large accelerators, with energies larger than 6 MeV, have the
additional possibility to use the γ-ray for PIGE analysis. For proton beams with an energy
above 6.5 MeV, there are large discrepancies between the measured cross-sections of Dyer et
al. [4.52] and Benhabiles-Mezhoud et al. [4.67]. A re-measurement of the 1635 keV γ-ray
production cross-sections at proton energies above 6.5 MeV could result in resolving the
discrepancies and provide a reliable cross-section for this γ-ray as well.
15
4.5.1.3. N(p,αγ1–0)12C : 4439 keV γ-ray

Thick-target yield measurements on natural nitrogen sample show that the intensity of the
4439 keV γ-ray is not negligible in comparison with the intensity of the 2313 keV γ-ray. In
spite of that, the abundance of 15N in natural nitrogen samples is only 0.368 % as compared to
the abundance of 14N, which is 99.632 %. Gorodetzky et al. [4.70] measured γ-ray production
cross-sections from 350 to 1910 keV, using a gas target enriched by 15N to 99.6 %. Rolfs and
Rodney [4.71] took measurements within the 150−2500 keV proton energy range and also
used a gas target. More recently, Imbriani et al. [4.72] repeated the cross-section
measurements in the 150−1910 keV proton energy range. The 15N enriched TiN target was
created by sputtering Ti onto Ta backings under nitrogen atmosphere that had been enriched
107
to 99.95% 15N. The reaction γ-rays were detected using a HPGe clover detector. The results
are presented in Fig. 4.34 for comparison.
FIG. 4.34. Differential γ-ray production cross-sections of the 4439 keV 15N(p,αγ1–0)12C resonance as a
function of proton energy, deduced from the work of a) [4.72], b) [4.71] and c) [4.70] at 90°. As the
former two cross-section data [4.72, 4.71] were given as total cross-sections, they were divided by 4π
to obtain differential cross-sections assuming isotropy.
The target of Gorodetzky et al. [4.70] was too thick (47 keV) for a precise cross-section
measurement. The resonances of the other two measurements are narrow. The background
between the resonances in the measurement made by Rolfs and Rodney [4.71] is quite high
due to a contamination in the target.

From the discussion in the previous section, it follows that only the cross-section values of
Imbriani et al. [4.72] can be recommended for further application. It should be noted,
however, that the purpose of these cross-section measurements was to provide data for
astrophysical use and the beam current was around 10 µA, which is much higher than usual
for the PIGE technique. Nevertheless, there are possibilities to use this γ-ray for PIGE
analysis, e.g. for biological applications with an external beam. Although the yield from this
reaction increases for proton beam energies greater than 2 MeV, the background increases
even more rapidly. For practical use, the optimum Ep region is about 1900−2100 keV [4.73].
108
4.5.2. Gamma-ray production cross-sections of nitrogen from deuteron-induced
nuclear reactions
Thick target yield measurements [4.18] showed that, contrary to what was observed for
proton-induced γ-ray yields, deuteron-induced γ-ray yields with nitrogen are very intense,
even at low deuteron energies, since there are several γ-rays in both the high and low energy
regions that produce high yields. The dominant nuclear reaction is the 14N(d,pγ)15N, with a Q
value of 8.609 MeV. Therefore the 15N nucleus decays from highly excited states. Before this
CRP, there was only one measurement of a thin target γ-ray production cross-section
available in the literature [4.74]. It was therefore imperative that new measurements were
performed and inlcuded in IBANDL in a timely manner.
14
4.5.2.1. N(d,pγ7–0)15N: 8310 keV γ-ray
The 8310 keV γ-ray originates from the 14N(d,pγ7–0)15N transition. Due to the spin ½ value of
the initial state, from which the γ-ray is emitted, the angular distribution of the 8310 keV γ-
ray is isotropic.
(a) New measurement

In the work of Csedreki et al. [4.75], the experiments were carried out using the 5 MV Van de
Graaff accelerator of MTA ATOMKI, Debrecen. Energy calibration of the beam was
determined with an uncertainty of 3 keV. The target chamber had several diaphragms in the
long entrance tube that formed a beam of 1 mm diameter, ending in a long Faraday cup and
was insulated from the rest of the beam pipe. This provided the possibility to measure
accumulated beam charge on the set-up, which was performed by a digital current integrator.
The γ-rays were detected using a 170 cm3 HPGe detector at 55o with respect to the beam
direction and 9.5 cm from the target, and it was shielded from γ-rays originating from the
Faraday cup. In order to determine the absolute efficiency of the γ-ray detector for every
nitrogen γ-ray, radioactive sources supplemented with narrow resonance reactions were used.
For Eγ < 3500 keV, calibrated radioactive sources (133Ba, 56Co, 60Co, 137Cs and 152Eu) were
placed at the exact position of the target. For Eγ > 3500 keV, the detector efficiency was
determined using γ-ray cascades from the 23Na(p,γ)24Mg and 27Al(p,γ)28Si reactions at 1417
and 992 keV resonance energies, respectively. The absolute double escape efficiency was also
calculated. With this knowledge, the contribution of the double escape line of the 8310 keV γ-
ray could be separated from the 7299 keV γ-line in the γ-ray production cross-section
calculations. The target chamber had an inlet at an angle of 135o for an ion-implanted silicon
particle detector with 500 µm active depth, 35 keV energy resolution and a 3 mm diameter
collimator positioned in front. The Si detector was used to detect backscattered particles
simultaneously with γ-rays from the target. The description of the experimental set-up,
accelerator energy calibration, particle detector solid angle determination and the absolute
efficiency determination of the γ-ray detector are described in [4.76]. The target was a thin
self-supporting silicon-nitride film with a nominal thickness of 200 ± 14 nm and an area of 5
2
× 5 mm , obtained from Norcada Inc. The number of target nuclides was determined with the
detection of backscattered deuterons from nitrogen, within the energy region where the
backscattering is considered Rutherford. The results were checked with the α-RBS technique.
Since nitrogen has complicated deuteron-induced PIGE spectra, only relatively intense γ-rays
were selected for yield measurements. The excitation function measurements were carried out
with 50 keV steps from 2.0 to 0.65 MeV and, at certain points, the measurement was repeated
twice. The obtained γ-ray production cross-sections for the 8310 keV γ-ray are presented in
Fig. 4.35 as a function of deuteron energy. The main sources of uncertainties are collected in
109
Table 4.19. The errors were added quadratically. The average uncertainty of the nitrogen γ-
ray production cross-sections is 5%.

In Fig. 4.35, the data of Bebber et al. [4.74], which is the only data found in the literature for
nitrogen, is presented for comparison. Gamma-yields of the 14N(d,pγ)15N reaction were
measured by bombarding thin Si3N4 films on silicon substrates prepared by plasma sputtering.
The thicknesses of the targets were determined by ellipsometry and the areal density of 14N
was obtained by using the nominal density of Si3N4. The areal density was confirmed by
nuclear reaction analysis. Ion charge was collected on the wings of a beam chopper, which
allowed the determination of the dose with a precision better than 2%. The γ-rays were
detected by a high-efficiency detection system (solid angle around 4π, photo-peak-efficiency
47% for 8 MeV γ-rays) including a 12′′ × 12′′ NaI(Tl) bore hole detector. Due to the summing
property of the detection system, practically no escape peaks appeared and the Compton
background was significantly reduced. This feature was especially advantageous in the case
of nitrogen, where the presence of escape peaks of high-energy γ-rays make the γ-spectrum
and its evaluation more complicated. Total cross-section curves for the γ-rays of 7299 and
8310 keV were obtained in the deuteron energy range of 0.5−1.5 MeV. In Fig. 4.35, following
convention, differential γ-ray cross-sections are obtained by dividing the given total cross-
section values by 4π. The errors of the cross-section values contain the inaccuracies that
resulted from the beam current and detection efficiency measurements as well as the error in
the determination of surface atomic density. Table 4.19 details the uncertainty budget of the
new measurement in comparison with the above literature data. Final uncertainties are in the
same range for both results. A rather large discrepancy between the two results can be seen.
TABLE 4.19. ESTIMATED UNCERTAINTY BUDGET FOR THE ABSOLUTE DIFFERENTIAL

CROSS-SECTION VALUES OF THE 8310 keV 14N(d,pγ7–0)15N FROM THE MEASUREMENTS
OF [4.75] AND [4.74].
Quantity [4.74] [4.75]
Beam energy spread (keV) 3.0 keV
Charge determination < 2% 3.0%
NT number of target nuclei per square centimeter 2.0–3.0% 2.5%
γ-ray peak area < 4.0%
γ-ray detector absolute efficiency ~2.0% 2.0%
Final average uncertainty 5.6% 5.0%
In Fig. 4.35, γ-ray production cross-section data of the new measurement are compared with
the data of Bebber et al. [4.74]. Following convention, differential γ-ray cross-sections are
obtained by dividing the total cross-section values given in [4.74] by 4π. A large discrepancy
between the two results can be observed.
110
FIG. 4.35. Differential γ-ray production cross-sections of the 8310 keV 14N(d,pγ7–0)15N resonance as a
function of deuteron energy, taken from the works of a) [4.75], b) [4.74].
A comparison of the integrated cross-section data with previously reported thick-target yields
confirms one of the differential cross-section curves from Fig. 4.35. For the 8310 keV γ-ray,
only the thick-target γ-yields of Elekes et al. [4.18] are available in the literature. Figure 4.36
illustrates that the data of Csedreki et al. [4.75] follow the tendency of the measured curve,
while Bebber’s data [4.74] gradually deviates from it. From a detailed comparison [4.77], it
was deduced that although both the calculated thick-target yields show large variances, most
of Csedreki’s data are within the 25% uncertainty band of the experimental thick-target curve.
Data from [4.74] has a systematic shift on the negative scale that gradually increases in
deviation. Furthermore, it was found that the evaluation of the large Doppler broadened 8310
keV peak could cause an additional uncertainty in the elimination of the background for both
thick and thin target γ-spectra. After removing data points charged with this possible
additional uncertainty, the data of Csedreki et al. remain within the 0−15 % uncertainty band,
while the data of Bebber et al. are 10−25 % lower than the measured thick-target yield.
111
FIG. 4.36. Thick-target yields of the 8310 keV 14N(d,pγ7–0)15N resonance as a function of deuteron
energy. Data obtained from a) measured yields [4.18], b) and c) calculated yields using cross-sections
of [4.75] and [4.74], respectively.

The γ-ray production cross-sections measured for the 8310 keV γ-ray from the 14N(d,pγ)15N
nuclear reaction by Csedreki et al. [4.75] during this CRP cover a larger deuteron energy
region and contain more data points than previous measurements. However, the large
discrepancy between the new and the old data, which are beyond the quoted experimental
uncertainties, still cannot be explained. To resolve the problem and be able to recommend
suitable cross-section data, an evaluation of the measured cross-sections is necessary. Further
validation, by comparing the results with measured thick-target yields, is suggested.
Additional differential cross-section measurements would also be useful.
14
4.5.2.2. N(d,pγ5–0)15N: 7299 keV γ-ray
The 7299 keV γ-ray originates from the 14N(d,pγ5–0)15N reaction. The initial state from which
the γ-ray is emitted has a 3/2 spin value. Therefore, an anisotropic angular distribution is
expected to be observed for this γ-ray.
(a) New measurement

In the work of Csedreki et al. [4.75], the differential cross-sections for the 7299 γ-ray and the
8310 keV γ-ray from the 14N(d,pγ)15N reaction were measured simultaneously. The 7299 keV
γ-ray is the strongest, however its full energy (FE) peak is contaminated by the double escape
(DE) peak of the 8310 keV γ-ray. For the determination of the cross-section of this γ-ray, the
above mentioned DE peak was subtracted based on the known absolute DE efficiency curve
of the HPGe detector. The resulting differential γ-ray production cross-sections and the
corresponding errors are presented in Fig. 4.37 as a function of beam energy.
112
The γ-ray production cross-sections of the 7299 keV γ-ray were also measured by Bebber et
al. [4.74]. These data are compared with the new measurements in Fig. 4.37. The advantage
of using a 4π γ-detector [4.74] was exploited in the case of this γ-ray, i.e. determining cross-
sections from the measured spectra without having to deal with the problem of disturbing
escape peaks. Moreover, the obtained total cross-sections are independent from anisotropic
angular distributions. However, there is a large discrepancy between the new and old results,
like in the case of the 8310 keV γ-ray. The total cross-sections of [4.74] were divided by 4π to
give differential cross-sections for comparison with the new data of [4.75]. The underlying
assumption is that this γ-line has negligible anisotropy within the experimental uncertainties.
Fig.
4.37. Differential γ-ray production cross-sections of the 7299 keV 14N(d,pγ5–0)15N resonance as a
function of deuteron energy. Data obtained from the works of a) [4.75] at 55° and b) [4.74].

The measurement of the γ-ray production cross-section for the 7299 keV γ-ray in the
14
N(d,pγ)15N nuclear reaction by Csedreki et al. [4.75] was compard with the only existing
data [4.74] prior to the CRP. Because these new cross-section data are ‘clean’ from the
contamination of the second escape peak of the 8310 keV γ-ray, they can not generally be
applied directly to PIGE analysis. Instead, the use of the 8310 keV γ-ray production cross-
section is recommended. Alternatively, the differential γ-ray producing cross-section of the
7299 + 8310 DE keV peak can be measured. The yield of this composed peak is much higher
than the yield of the 8310 keV γ-line, however the determination of the ratio of the DE/FE
peak depends strongly on the detector geometry.
113
14
4.5.2.3. N(d,pγ6–1)15N: 2297 keV γ-ray
The initial state from which the 2297 keV γ-ray is emitted has 7/2 spin value. Therefore the
angular distribution of the 2297 keV γ-ray is expected to be anisotropic.
(a) New measurement

The cross-section for the production of the 2297 keV γ-ray from the 14N(d,pγ6–1)15N reaction
was measured by [4.75] and the results are plotted in Fig. 4.38. No previous measurement
exists in the literature.
FIG. 4.38. Differential γ-ray production cross-section of the 2297 keV 14N(d,pγ6–1)15N resonance
measured at an angle of 55° as a function of deuteron energy. Data obtained from the work of
a) [4.75].
(b) Assessment
The integrated cross-section data were compared with previously reported thick-target yields
[4.18]. From the result presented in Fig. 4.39, one can conclude that the calculated yield is
systematically larger than the measured thick-target yield. However, this deviation remains
within the 25% uncertainty band.
114
FIG. 4.39. Thick-target yields of the 2297 keV 14N(d,pγ6–1)15N resonance as a function of deuteron
energy. Data obtained from thick target yields a) measured at 135° [4.18] and b) calculated from the
differential cross-sections of [4.75] measured at 55°.

The differential γ-ray production cross-section values of the 2297 keV γ-ray are about 70% of
the corresponding values of the 8310 keV γ-ray when induced by deuteron beam with energy
around 2 MeV. However, there are some advantages in using this γ-ray for PIGE analysis, e.g.
its Doppler broadening is smaller due to its lower energy, which makes it easier to subtract
the background. The determination of the efficiency of the γ-detector is also straightforward,
since, at lower γ-energy, it can be performed solely with radioactive sources without the need
to use nuclear reactions.
14
4.5.2.4. N(d,pγ4–1)15N: 1885 keV γ-ray
For the 1885 keV γ-ray of the 14N(d,pγ4–1)15N reaction, the initial state emitting the γ-ray has a
5/2 spin value, which indicates that an anisotropic angular distribution is expected for this γ-
ray. The γ-ray is also Doppler broadened due to the 18 fs lifetime of the initial state.
(a) New measurement

Csedreki et al. [4.75] also measured these cross-sections for γ-rays and results are plotted in
Fig. 4.40. No other measurement can be found in the literature.
115
FIG. 4.40. Differential γ-ray production cross-sections of the 1885 keV 14N(d,pγ4–1)15N resonance
measured at an angle of 55° as a function of deuteron energy. Data obtained from a) [4.75].
(b) Assessment
The resulting differential γ-ray production cross-sections and corresponding errors are
presented in Fig. 4.40 as a function of incident beam energy. The energy dependence of the
obtained cross-sections shows deviations from the direct reaction mechanism. This is clearly
observed for the 1885 keV γ-ray, where a pronounced maximum is seen in the cross-section
function in the studied energy interval.
The integrated cross-section data was compared with previously reported thick-target yields
[4.18]. From the result presented in Fig. 4.41, it can be concluded that the calculated yield is
systematically larger than the measured thick-target yield. However, this deviation is not higher
than 15%.
116
FIG. 4.41. Thick target yields of the 1885 keV 14N(d,pγ4–1)15N resonance as a function of deuteron
energy. Data obtained from a) measured yields at 135° [4.18] and b) calculated yields from the
differential cross-sections of [4.75] measured at angle of 55°.

The advantage of using the resulting cross-section values for PIGE analysis is that for the
determination of the γ-detector efficiency, no use of nuclear reaction measurement is required.
It is recommended that the 1885 keV γ-ray be used, e.g. in simultaneous measurement of
nitrogen and oxygen containing samples.
One additional remark is that the calculated thick-target yields for all four γ-rays of the
14
N(d,pγ4–1)15N reaction are higher than the measured γ-yields of Elekes et al [4.18]. In the
absence of any other experimental data, additional measurements are recommended for both
thin and thick targets in order to properly validate the results.
4.6. FLUORINE
4.6.1. Gamma-ray production cross-sections of fluorine from proton-induced nuclear

reactions
19
4.6.1.1. F(p,p′γ)19F: 110 keV and 197 keV γ-rays

Four laboratories, RBI, LABEC-INFN, CTN/IST, and CMAM-UAM, took responsibility for
measuring cross-sections for PIGE emission produced by protons on 19F. Two different γ-rays
were considered: the 19F(p,pγ1−0)19F (Eγ = 110 keV) and the 19F(p,pγ2−0)19F (Eγ = 197 keV),
while two other γ-rays, the 19F(p,pγ3−1)19F (Eγ = 1236 keV) and 19F(p,pγ4 − 1 + 5 − 2)19F (Eγ =
1349/1357 keV), have been measured so far only by the group at RBI [4.79]. The yield of the
117
latter two γ-rays was known to be significantly lower than those for the first two. Since the
first two γ-rays are more suitable for PIGE analysis, they were the only ones included in the
list of planned measurements for the CRP.
The 110 keV γ-ray has an isotropic angular distribution (due to the spin -1/2 of the initial first
excited state) while the 197 keV γ-ray is expected to have an anisotropic angular distribution
(spin -5/2 of the initial second excited state of 19F). The cross-sections have been measured
for proton energies spanning from 815 keV to 4415 keV. Each group covered a different
energy range with a common overlap between 2133 and 3022 keV. As a result, there are at
least two measurements for any given part of the explored region. The laboratories used
different experimental conditions, which are summarized in Table 4.20 and have produced, as
a consequence, different uncertainty budgets, which are summarized in Table 4.21.
RBI measured target areal densities ratios by 1.8 MeV a BS obtaining N19F/NAu = 47 ± 1,
N24Mg/NAu = 18.3 ± 0.4 N25Mg/NAu = 2.32 ± 0.05. The detector efficiency was measured
by placing calibrated sources of 60Co, 133Ba, 137Cs and 152Eu in the exact target position.
Afterward, the absolute efficiency at 110, 197, 1236 and 1349 keV was obtained from a third
degree polynomial fit in the argument 1/Eγ. To avoid absolute collected charge measurements,
the RBI, LABEC-INFN and CTN/IST groups extracted PIGE differential cross-sections,
relative to the known values of the proton Rutherford backscattering cross-sections and the
known quantity of high Z element (Au or Ag) present in the target. In this case, the particle
detector was a Si surface barrier whose solid angle was determined by the RBS of 1.8 MeV
protons of thick single element targets (Ni, Sn, Ta, Au). The energy calibration was
performed using the resonance at 991.88 ± 0.04 keV of the 27Al(p,γ)28Si reaction and the
7
Li(p,n)8Be reaction threshold at 1880.44 ± 0.02 keV. After calibration, the proton beam energy
was known with an accuracy of 0.06%.
LABEC-INFN used the same targets as the CTN/IST, i.e. a NaF film evaporated on a self-
supporting Ag film. The target thickness was NaF = 34.9 ± 1.5 µg/cm2 and Ag = 51.0 ± 1.8
µg/cm2 and the atomic ratio was F/Ag = 1.73 ± 0.05, as measured by 2.0 MeV α scattering.
They performed the same absolute efficiency calculation as RBI, using calibrated sources of
133
Ba and 152Eu. For normalization purposes and to avoid absolute charge measurement,
elastic scattered protons were collected simultaneously using a Si-PIN diode detector with a
10 mm2 area, 300 mm thickness and 13 keV FWHM energy resolution, placed at a backward
angle of 150° and collimated by a 1 × 13 mm2 vertical slit. The subtended solid angle was
3.37 ± 0.08 msr, while the spread in the scattering angle was 1.0° due to beam size on the
target and detector finite aperture. The proton beam energy was calibrated using an aluminum
thick target and the resonances at 991.86 keV and 1683.57 keV in the (p,γ) and (p,p′γ) reactions,
respectively, on 27Al. The proton beam energy was also calibrated using a native silicon oxide thin
target and the resonance at 3470 keV in the proton elastic scattering from 16O. After the
calibration, the proton beam energy was known with an accuracy of 0.15%.
The CTN/IST group produced the NaF targets that were also used by LABEC-INFN.
Analysis by a particle RBS at 165° and incident energies of 1600 and 1800 keV gave the
following thicknesses: F = 5.1 × 1017 at/cm2, Ag = 3.2 × 1017 at/cm2 with an atomic ratio
F/Ag = 1.59, about 9% lower than the Florence independent measurements average. The
absolute efficiency versus γ-ray energy was obtained by measuring the yields of γ-rays
produced by ion sources (133Ba and 153Eu) calibrated in activity as well as by Monte Carlo
simulations using GEANT and PENELOPE. For normalization purposes, scattered protons
were detected at 157° to the incident beam by Camberra PIPS detectors with an active area of
50 mm2 and a depletion layer of 100 mm. The detector was located at 88 mm from the target,
118
following Cornell geometry, with solid angles defined by apertures of 6.00 ± 0.05 mm. Proton
energy was calibrated using the 1645 and 1931 keV resonances of the 23Na(p,p′γ)23Na
reaction and the 3470 keV resonance of the 16O(p,p)16O reaction.
119
TABLE 4.20. COMPARISON OF EXPERIMENTAL CONDITIONS OF THE FOUR LABORATORIES MEASURING THE CROSS-SECTIONS OF THE
PROTON-INDUCED 19F REACTIONS
Parameters RBI LABEC-INFN CTN/IST CMAM-UAM
Gamma-detector 1.30% 0.252 ± 0.018% Not explicitly quoted 0.40%

absolute-efficiency at ± 7–10% (relative) ± 4.6% (relative)
110KeV
Gamma-detector 1.04% 0.213 ± 0.013% Not explicitly quoted 0.28%
absolute-efficiency at ± 7–10% (relative) ± 1.9% (relative)
197KeV
Gamma-detector 135° 45° 130° 135°
angle
Gamma-detector 6.6 cm 18.6 cm 5.6 cm 21.1 cm
distance to target
Target MICROMATTER™ MgF2 54.1 NaF evaporated on Ag film NaF evaporated on Ag film Various LiF targets (see text)
µg/cm2 on 6.3 mm Mylar foil + self-supporting. self-supporting.
4 nm Au layer evaporated on NaF 34.9 ± 1.5 µg/cm2, NaF 34.9 ± 1.5 µg/cm2,
top Ag 51.0 ± 1.8 µg/cm2 Ag 51.0 ± 1.8 µg/cm2
Beam size 4 mm ∅ 1.0 × 1.0 mm2 Not reported 1.5–3 mm ∅
Energy range 1950–3050 keV 2463–4101 keV 2144–4059 keV 815–4415 keV
Beam Currents 10–20 nA 70–140 nA < 200 nA 150–300 nA
Particle detector solid 7.7 +/- 0.2 msr 3.37 ± 0.08 msr 3.65 ± 0.06 msr Not used
angle
120
TABLE 4.21. RELATIVE ERRORS AND UNCERTAINTY BUDGET FOR THE FOUR LABORATORIES MEASURING THE CROSS-SECTIONS OF
THE PROTON-INDUCED 19F REACTIONS
Parameters RBI LABEC-INFN CTN/IST CMAM-UAM
γ-detector ± 7–10% ± 7% ± 5% ± 1.9–4.6%

absolute-efficiency
Particle detector solid ± 2% ± 2.4% ± 2.5%
angle
γ-ray peak area < ± 1% (110–197 KeV) ± 2–3% ± 1.4% ± 1%
< ± 4% (1236–1349 + 1357 KeV (resonances)
at Ep > 2.2MeV) ± 3.2%
< ± 22% (1236 at Ep < 2.MeV) (between resonances)
< ± 36% (1349 + 1357 KeV
at Ep < 2.2MeV)
Scattered particles peak ± 1% ± 1% ± 1%
area
Target ± 5% (MgF2) ± 4.3% (NaF) ± 4.3% (NaF) ± 20%
Thickness ± 3.5% (Au) ± 3.5% (Au)
(as measured at LABEC-
INFN)
Atomic ratio ± 2% ± 2% ± 3%
Rutherford cross-section ± 1% ± 1%
Colected charge ± 2.5%
Cumulative error ± 10% (110 and 197 KeV) ± 8–8.3% ± 6.5–7.1% ± 20–21%
±10–24% (1236keV)
±10–38% (1349 +1357keV)
121
The CMAM-UAM lab measured the absolute detector efficiency using the same procedure
and sources as the RBI group. However, due to the fact that the entrance window of the REGe
Canberra detector was a polymer, the efficiency had a maximum value at the 80.9979 keV of
the 133Ba line. Various targets were used to measure in different energy regions: LiF/Ag/C
(nominal thickness: 13µg/cm2 LiF + 40µg/cm2 Ag), Au/LiF/C (50µg/cm2 Au + 20 µg/cm2 LiF
+ 15 µg/cm2 C) LiF/Ag/C (20 µg/cm2 LiF + 40 µg/cm2 Ag + 30 µg/cm2 C). The first target
was supplied by CTN/IST and the other two were supplied by the Laboratori Nazionali di
Legnaro (INFN-LNL). The target thickness measured by RBS and γ-yield, could not be
accurately reproduced and the estimated systematic error is ± 20%. With such
ahighsystematic error dominating the cumulative error, this dataset was not suitable for fixing
absolute values. However, it remains useful for studying the structure of the excitation
function, since it covers the largest energy range among all the measured datasets. The proton
beam energy was calibrated using the resonances at 991.86 keV and 1316.83 keV in the
27
Al(p,γ)28Si reaction as well as the non-resonant reactions 55Mn(p,p)55Mn, natSn(p,p)natSn, and
11
B(p,α)8Be. After the calibration, the proton beam energy was known with an accuracy of
0.07%.
(b) Assessment
The data from Chiari et al. (LABEC-INFN) [4.78], Jesus et al. (CTN/IST) [4.10], and
Zamboni et al. (RBI) [4.79] are shown in Figs. 4.42 and 4.43. All the data are available in
IBANDL, however, at the present stage, only the latter has been published. While the RBI
and CTN/IST energy scales coincide, the LABEC-INFN data and the CMAM-UAM data
needed to be shifted up by 10 keV. After the correction, the 110 keV data are all within ± 8%
at the energy around 3200 keV and therefore fully within the evaluated cumulative errors.
Below 3 MeV, the data are in close agreement. Above 3 MeV, only two data sets can be
compared, which differ in shape in some energy regions (3.4–3.5 and 3.7–3.9 MeV). The
overall agreement, however, is within quoted uncertainties.
FIG. 4.42. Differential cross-sections for the 110KeV 19F(p,p’γ)19F resonance as a function of proton
energy. Data measured during the CRP by a) Jesus et al. [4.78], b) [4.79] and c) Chiari et al. [4.78]
122
The situation is different for the 197 keV data, where the difference between datasets is a
factor of 2 outside the quoted cumulative errors. Since the two yields have been measured
from the same spectra, such a different behavior in the 197 keV case with respect to the 110
keV case requires an explanation. The only factors which are different for the 110 keV and
the 197 keV line are: the γ-ray angular anisotropy (according to Allegro et al. [4.80] the
anisotropy of the 197 keV γ-line can be very strong), relative detector efficiency, peak
background subtraction and the possible presence of background peaks coming from the F
presence in the experimental setup (even at the trace level such as in Ta collimators) or beam
stoppers. The pile-up might affect the two peaks differently but it has been maintained at a
low level. A systematic error of 25% or even 50% in the efficiency of the 197 keV peak
would be unexpected given that the two energies are very similar and the experimental
efficiency values should not be that different.
FIG. 4.43. Differential cross-sections for 197 keV 19F(p,p′γ)19F resonance as a function of proton
energy. Data as measured within the CRP by a) Jesus et al. (130°) [4.78], b) [4.79] (135°) and
c) Chiari et al. (45°) [4.78].
(c) Data in literature

Already in 1986, Chaudri [4.81] observed that “in spite of the inherent advantage of the
F(p,p′γ)F, its use for analytical and depth profiling purposes has been rather limited, probably
due to the non-availability of accurate cross-section data”. The situation represented in Fig.
4.44, where all the published data [4.5, 4.6, 4.9, 4.79, 4.81–4.83] are compared to the CRP
produced data, still supports Chaudri’s statement. It is worthwhile noting that a correlation
analysis performed on the 110 keV gamma cross-sections and interpolated at all the measured
energies gives Pearson correlation coefficients (Table 4.22) that are high in all cases except
for the data of Ranken [4.84]. The Pearson coefficient between two variables (in our case the
two cross-section functions that are compared) depends on the standard deviation of the two
variables and on their covariance, but is independent of the measure units. A value of 1
123
indicates perfect correlation and 0 indicates no correlation at all. In other words, the relative
heights of structures in the excitation functions are constant from one set to the other but the
absolute values are evidently affected in many cases by large systematic errors.
The work of Boni et al. [4.5] is one of the most complete in regards to the number of elements
investigated. A LiF target of thickness 1000 Å deposited on a Nuclepore™ film was used.
Precision is quoted at 15% for the absolute values of their differential cross-sections, but an
uncertainty budget is not provided. It is reported that the germanium detector was calibrated
with 226Ra. The lowest line of 226Ra, however, is at 186 KeV, which means that the efficiency
for the 110 keV F line had to be extrapolated. The Faraday cup was rather distant from the
target, which might have enhanced transmission problems due to the multiple scattering of the
beam in the target. Several authors have found that their data, in cases other than F, can differ
by a factor of up to 5 from those of Boni et al.
The work of Ranken et al. [4.84] was a pioneer one. The CaF2 targets were evaporated on
tungsten and supported by a 3.175 mm aluminium plate. Having set the detector at 45° to the
beam direction, the photons were partially absorbed by the target backing. Protons stop in the
target and the charge was collected through a calibrated integrator. Photons were detected by
a 1” × 1” NaI scintillator. Both the detector efficiency and the photon absorption in the target
were computed from tables and diagrams. An accuracy of 10–15% is quoted for the absolute
values of the total cross-sections. No uncertainty budget is detailed. The consistency of the
data is claimed to have come from the comparison with the 110 keV data of Barnes [4.85] at
1.70 MeV. While the NaI absolute efficiency is good, the resolution is not comparable to that
of a germanium detector and the background/signal ratio is much higher. A cumulative error
of 10–15% compared to a cumulative error of 7–8% (CRP data) might be optimistic.
Chaudri [4.44] covered a large energy range from 0.5 to 4.3 MeV in steps of 20 keV. He used
a CaF2 target on a carbon foil and measured the thickness with RBS (probably protons), with
an uncertainty < 5%. He used a large volume NaI scintillator at 125°, which covered an
angular range of 63.5°. It is demonstrated that the yield obtained in his set-up from the
anisotropic 197 keV emission is only 3% different from the yield obtained from an isotropic
emission. The detector efficiencies are calculated to be 4.2% ± 0.27% at 110 keV and 3.71%
± 0.24% at 197 keV. No detail is given on this calculation and no check with a calibrated
source has been done. Given the use of a NaI scintillator, the same considerations as for the
work of Ranken et al. [4.1] hold. It is worth noting that Chaudri´s data are not available in
tabular form and have been digitized from the publication plots (in log scale), which
introduces a further source of error.
124
FIG. 4.44. Differential cross-sections for the 110 keV 19F(p,p′γ)19F resonance (TOP) and 197 keV
19
F(p,p′γ)19F (BOTTOM) as a function of proton energy. Data from all the published and CRP
produced datasets: a) Jesus et al. at 130° [4.78], b) [4.79] at 135°, c) Chiari et al. at 45° [4.78],
d) CMAM-UAM at 135°, e) [4.82] at 130°, f) [4.6] at 135°, g) [4.83] at 90°, h) [4.5] at 90°, i) [4.84]
at 45° and j) [4.81].
125
TABLE 4.22. THE PEARSON CORRELATION COEFFICIENTS OF THE AVAILABLE SETS OF DATA, INTERPOLATED AT ALL MEASURED
ENERGIES FOR THE 110 keV 19F(p,p′γ)19F RESONANCE
CHAUDRI CACIOLLI CTN/IST2000 CTN/IST BONI RANKEN RBI GRAMBOLE LABEC-INFN CMAM-UAM
CHAUDRI 0.993 0.992 0.956 0.986 0.995 0.994 0.990 0.988 0.992
CACIOLLI 0.988 0.959 0.944 0.988 0.960 0.975 0.992
CTN/IST2000 0.987 0.986 0.987 0.988 0.993 0.981 0.983 0.992
CTN/IST 0.956 0.951 0.981 0.928 0.942 0.971 0.953 0.968 0.971
BONI 0.984 0.960 0.989 0.949 0.987 0.988 0.989 0.984 0.966
RANKEN 0.863 0.604 0.973 0.758 0.921 0.992 0.993 0.980 0.984
RBI 0.994 0.949 0.997 0.978 0.983 0.938 0.987 0.985 0.983
GRAMBOLE 0.966 0.945 0.969 0.981 0.978 0.985 0.986 0.990
LABEC-INFN 0.962 0.978 0.969 0.950 0.962 0.704 0.978 0.987 0.984
CMAM-UAM 0.976 0.961 0.992 0.947 0.941 0.825 0.982 0.969 0.967
126
Grambole et al. [4.83] used a 20 nm CaF2 target on a Ta backing and detected low energy
photons with a Ge(Li) X-ray detector with a 16 mm diameter and a 7 mm thickness. The
detector was placed at 90° with respect to the beam direction, with the target tilted at 45°. The
excitation curves range from 0.69 to 2.94 MeV and are given in relative units. No errors are
quoted. Although the numerical values are quite close to the differential cross-section values
of the other authors, the use of relative units is indicative of possible systematic errors. The
Grambole data below 1600 keV appear shifted by about -10 keV.
The two other published sets of data [4.82, 4.6] are much more recent and were produced by
groups taking part in the CRP. Both were obtained using a set-up very similar to the set-up in
the CRP data collection. Both produced an uncertainty budget, with a cumulative,
conservative error of ± 15% for Caciolli et al. [4.6] and 9.6–12.6 % for Jesus et al. [4.82].
As previously stated, with the exclusion of [4.84] for the 110 keV channel, there is a clear
correlation between the different data sets. Considering that, due to the coordination that the
CRP has assured, the data produced by Zamboni et al. [4.79], Jesus et al. [4.78] and Chiari et
al. [4.78] for the 110 keV case are accurate, the previous data can be renormalized to these
results, as is seen in Table 4.23.
TABLE 4.23. LINEAR COEFFICIENTS OF THE AVAILABLE SETS OF DATA,

INTERPOLATED AT ALL MEASURED ENERGIES.
Data Set Normalization Note for 110 keV Normalization Note for 197 keV
to 110 KeV to 197 KeV
CTN/IST N/A N/A 1.3–1.7 Dependent on energy
and outside errors
CTN/IST2000 0.85 Within the errors 1.4 Outside the errors
Caciolli 0.89 Within the errors 1.15 Within the errors
CMAM- 0.80 Outside the errors 1.08–1.15 Depending on energy
UAM quoted for the but within the errors
reference data. quoted for the
reference data
Grambole 0.95 Within the errors 1.0 Within the errors
Boni 0.6–0.7 Dependent on 0.9 Within the errors
energy and outside
errors
Chaudri 0.57–0.72 Dependent on 0.78 Outside errors
energy and outside
errors
Ranken N/A Poor correlation 1.3 Outside errors
In the case of the 197 keV line, there is no agreement between the CRP produced data and the
data-spread gets wider with increasing energy, up to a maximum of ± 25%. It is surprising to
note that the two series measured at CTN/IST (Jesus et al. [4.78] and [4.82]) match for the
110 keV and the 197 keV case but only the 110 keV measurements agree with the rest of the
CRP data. Since the data of Zamboni et al. and Chiari et al. [4.78] for the 197 keV γ-ray agree
in the common energy range, it is assumed that they are accurate. Therefore, the re-
normalization process, as reported in Table 4.23, is repeated. The data of Chaudri [4.81] and
Ranken [4.84] do not fit with the rest because their discrepancy is energy dependent for the
110 keV case. Except for the case of Grambole [4.83], where it is assumed that their relative
127
units are indeed mb/sr, the normalization coefficients are remarkably different for the two
cases, 110 keV and 197 keV γ-lines, which is probably due to an unexpectedly large
difference in the measurement of the absolute detector efficiency for the two photons.
The data included in this detailed analysis had a cut-off date of December 2016. Since then,
there has been an additional publication of new data from the CMAM group [4.78]. These
data were not considered in our assessment; however they are available in IBANDL.
4.6.2. Thick target yields
In addition to the normalization procedure, differential cross-section data were compared to

the existing measured thick-target yields. For that purpose, an integration routine developed
by Lagoyannis [4.85] was used.
19
4.6.2.1. F(p,pγ1–0)19F: 110 keV γ-ray
For the 110 keV γ-ray, only three experimental datasets exist for thick-target yield (Kenny
[4.86], Antilla [4.1], and Kiss [4.2]) Fig. 4.45 shows thick-target yields calculated using the
experimental data measured in the present CRP and the available thick-target yield data [4.86,
4.1, 4.2]. In order to compare the new measured differential cross-sections with the thick-
target yield data sets, the integrated cross-section data were normalized to the data of Kenny
et al. [4.86] at the energy of 2.5 MeV. The normalization was performed by adding a constant
number to the integral to take into account that the measured differential cross-sections do not
extend to zero energy. It is worth noting that it is not possible to normalize the integrated
differential cross-section data to the thick-target yields of Antilla [4.1] and Kiss [4.2], since
the latter’s absolute value is smaller than the integrated cross-sections and would therefore
require subtracting a constant instead of adding it, which is not physically possible.
128
FIG. 4.45. Comparison of thick target yields for the 110 keV 19F(p,p′γ)19F resonance as a function of
energy. Calculated yields are obtained with cross-section data from (a) Chiari et al. [4.78], (d) [4.79],
(e) Jesus et al. [4.78], while measured yields are taken from (b) [4.86], (c) CTN/IST [4.2] (f) [4.1].
In Fig. 4.46 thick-target yields are compared just as in Fig. 4.45, but with the addition of
thick-target yields calculated using old experimental data from IBANDL. To compare new
and old measured differential cross-sections with thick-target yield datasets, the integrated
data were normalized to Kenny’s data [4.86] at 2.5 MeV, by adding a constant to take into
account that the measured differential cross-sections do not extend to zero energy. Again, this
normalization procedure cannot be applied to the thick-target yields of Kiss [4.2] and Antilla
[4.1] for the same reason mentioned above.
129
FIG. 4.46. Comparison of the thick-target yields for the 110 keV 19F(p,p′γ)19F resonance as a function
of energy. Calculated yields are obtained with cross-section data from (a) Chiari et al. [4.78],
(d) [4.79], (e) Jesus et al. [4.78], and (g) [4.6] and measured yields are taken from (b) [4.86],
(c) [4.2], and (f) [4.1].
19
4.6.2.2. F(p,pγ2–0)19F: 197 keV γ-ray
For the 197 keV line, there are four experimental data sets for thick-target yields (Kenny
[4.86], Kiss [4.2], Antilla [4.1] and Savidou [4.17]), which are compared with calculated
yields obtained from the various experimental differential cross-section data (new and old)
available in the literature and shown in Fig. 4.47. The same normalization procedure was
applied as in the case of the 110 keV γ-ray described in the previous section. The data of
Savidou at al. [4.17] suffer from the same defficiency as those of Kiss [4.2] and Antilla [4.1],
as described above for the 110 keV γ line. For this reason, the comparison is limited to the
data of Kenny et al. [4.86]. All datasets agree in the shape of the yield curve produced up to
an energy of 2.5 MeV. Futhermore, the previous low energy data of Jesus et al. [4.82],
extending up to 2.7 MeV, have been validated by an independent benchmark performed by
Matheus et al. [4.87]. At energies above this low energy region, however, the datasets
diverge. This is expected, given the divergence of the differential cross-section data observed
in Fig. 4.44. In the absence of any reliable experimental thick-target yields, it is not possible
to draw any definite conclusions for this energy region.
130
FIG. 4.47. Comparison of the thick target yields for the 197 keV 19F(p,p′γ)19F resonance as a function
of energy. Calculated yields are obtained from cross-section data of (b) Jesus et al. [4.78], (c) Chiari
et al. [4.78], (d) [4.82], (e) [4.84], (f) [4.79], and (j) [4.5]. Measured yields are taken from (a) [4.17],
(g) [4.86], (h) [4.1], (i) [4.17].
(d) Recommended data

PIGE bulk analysis requires knowing the gamma emission cross-section from the beam
proton energy down to zero in order to calculate the integral that provides elemental
concentrations.
From the above analysis, it is clear that the 110 keV γ-line is best suited for bulk analysis. It is
recommended to use any of the three new datasets that were measured during the CRP up to
3.2 MeV. At higher energies, the cross-sections have uncertainties of approximately 8%.
Another possibility is to use a combination of datasets from Chiari et al. [4.78], Zamboni et
al. [4.79], and Jesus et al. [4.78], that were produced in the CRP and combine them with the
two datasets at low energy, [4.82] and [4.6], using the normalization constants shown in Table
4.23. According to Table 4.23, the low energy data [4.82] should be multiplied by a factor
0.85 while the data at high energy [4.6] should be multiplied by a factor 0.89. The
recommended cross-section is shown in Fig. 4.48. This will cover a wide energy range,
corresponding to the most used PIXE energies and extending to 5.7 MeV for deeper bulk
analysis.
131
FIG. 4.48. Differential cross-sections for the 110 keV 19F(p,p′γ)19F resonance, including data from the
CRP and previous work [4.82, 4.6] as a function of proton energy. Data from a) Jesus et al. [4.78],
b) [4.79], c) Chiari et al. [4.78], d) [4.82] and e) [4.6].
For the 197 keV γ-line, the comparison of differential cross-section data and thick-target
yields have shown divergencies at energies above 3 MeV that cannot be explained. Therefore,
no recommendation can be made for those high energies. At lower energies up to 3 MeV, the
data are in better agreement; in particulary the data of Jesus at al. [4.82] have already been
benchmarked, which enables it to be used for bulk analysis. Nevertheless, to resolve the
above-mentioned discrepancies, it is recommended that an evaluation of the differential cross-
section data and additional measurements of thick-target yields are performed, over the whole
energy range, in order to validate the evaluated data.
4.7. SODIUM
4.7.1. Gamma-ray production cross-sections of sodium from proton-induced nuclear

reactions

In the framework of the CRP, differential cross-sections for proton-induced γ-ray emission
from 23Na(p,p′γ)23Na (Eγ = 441 and 1636 keV) and 23Na(p,α′γ)20Ne (Eγ = 1634 keV) have
been measured by five different laboratories LABEC-INFN, CTN/IST, ATOMKI, NSTRI and
RBI. Measurements were performed in the energy range between 1.0 to 4.1 MeV and at
several angles in the laboratory frame of reference, since these γ-rays have anisotropic angular
distributions.
The measurements at the HVEE 3 MV Tandetron accelerator at the LABEC-INFN laboratory

in Florence were done using + 30° beam line chamber. The groups at LABEC-INFN and
CTN/IST have used NaF films evaporated on a self-supporting Ag film. Targets were
produced at CTN/IST. The ratio between Na and Ag atoms was determined using 2 MeV He
132
RBS and was found to be 1.79 ± 0.04. The LABEC-INFN lab used three HPGe detectors to
detect γ-rays at 90°, 45° and 0°. They performed the same absolute efficiency calculation as
RBI, using calibrated sources of 133Ba and 152Eu. To avoid absolute charge measurement and
for the normalization procedure, elastic scattered protons were collected simultaneously using
a Si-PIN diode detector with 13 keV energy resolution, area 10 mm2 and thickness 300 mm.
Detector was placed at 150° and collimated by a 1 × 13 mm2 vertical slit; the subtended solid
angle was 3.37 ± 0.08 msr. The proton beam energy was calibrated using an aluminum thick
target and the resonances at 991.86 keV and 1683.57 keV, in the (p,γ) and (p,p′γ) reactions on
27
Al respectively, as well as a native silicon oxide thin-target on the resonances at 3470 keV
in the proton elastic scattering from 16O. After calibration, the proton beam energy was known
with an accuracy of 0.1%.
RBI measurements were performed using a conventional ion beam analysis scattering
chamber, installed at the -15° line of the 6 MV Tandem Van de Graaff accelerator. The proton
beam energy was calibrated using the 991.86 keV resonance of the 27Al(p,γ)28Si reaction and
the 7Li(p,n)7Be reaction threshold at 1880.44 keV on thick aluminum and lithium targets
respectively. After calibration, a beam energy accuracy of 0.06% was determined.
Micromatter thin reference standard (53.7 µg/cm2 of NaCl deposited on 6.3 µm Mylar) was
used for the measurements. For normalization purposes, a 3.5 nm Au layer was evaporated on
top of the target. In order to determine the ratio r between the Na and Au areal densities,
assuming Rutherford cross-sections, 1.8 MeV α particles were used. The ratio r was measured
to be 28.4 ± 0.7.
At ATOMKI, the measurements were carried out using the 5 MV Van de Graaff accelerator.
After energy calibration with the 991.86 keV resonance of the 27Al(p,γ)28Si nuclear reaction,
beam energy uncertainty was 1.5 keV. The absolute HPGe detector efficiency was determined
using radioactive sources (60Co and 152Eu, calibrated by the Hungarian National Office of
Measures) that were placed at target position. The target used for the measurements was
produced at CTN/IST (NaCl thick 41 µg/cm2 and evaporated on Ag foil). The ratio r between
Na and Ag was determined from the 1.5 MeV He RBS to be 2.52 ± 0.08.
At CTN/IST, measurements were performed using the 3.0 MV Tandetron accelerator and a
scattering chamber electrically insulated from the rest of the beamline. Proton beam energy
was calibrated with an accuracy of 0.1% using the 1645 keV and 1930.7 keV resonances of
the 23Na(p,p′γ)23Na reaction and by the 3470 keV resonance of the 16O(p,p)16O reaction. The
NaF target evaporated on thin Ag was used for the measurement and was analysed by 1.6 and
1.8 MeV RBS at 165° scattering angle. Atomic ratio of Na to Ag was found to be 1.60 ± 0.05.
The JPGe detector absolute efficiency was obtained by measuring the yields of γ-rays
produced by radioactive sources (133Ba and 153Eu) calibrated in activity as well as by Monte
Carlo simulations using GEANT and PENELOPE codes.
The proton beam from the 3.0 MV Van de Graaff of NSTRI was used for the measurements.
Proton energy was calibrated by the resonance at 991.86 keV of 27Al(p,γ)28Si and the
7
Li(p,n)7Be reaction threshold at 1880.44 keV. After calibration, the proton beam energy was
known to an accuracy < 0.2%. Thin films of sodium chloride were evaporated onto thin flims
of self-supporting Ag. The targets obtained this way were analysed by alpha RBS at 165° for
alpha particle incident energies of 2.0 MeV. Thicknesses of 1.9 × 1017 at/cm2 and 4.8 × 1017
at/cm2 were obtained for Na and Ag, respectively.
133
More details about the experimental set-ups used by different groups can be found in Table
4.24 and in [4.88]. The uncertainty budget, as estimated for the measurements performed in
the different laboratories, can be found in Table 4.25. It includes the statistical errors from the
peak areas with subtracted background, and different systematic errors. From there, it can be
estimated that an overall uncertainty of 7–8% can be expected for the cross-section of the 441
keV γ-ray line from the 23Na(p,p′γ)23Na reaction, and of 8–12% for the cross-section of the
1634–1636 keV γ-ray lines from the 23Na(p,α′γ)20Ne and 23Na(p,pγ2–1)23Na reactions.
134
TABLE 4.24. EXPERIMENTAL CONDITIONS USED IN THE FIVE LABORATORIES MEASURING THE DIFFERENTIAL CROSS-SECTIONS OF THE
23
Na(p,p′γ)23Na: 441 keV & 1636 keV γ-rays AND 23Na(p,α′γ)20Ne: 1634 keV γ-ray
Experimental Laboratory
Condition
LABEC-INFN RBI ATOMKI CTN/IST NSTRI
γ-ray detector (1.88 ± 0.11) × 10-3 (5.40 ± 0.38) × 10-3 (4.40 ± 0.13) × 10-3 (1.07 ± 0.05) × 10-2 (1.58 ± 0.08) × 10-4
absolute efficiency
(1.17 ± 0.07) × 10-3
at 441 keV
(6.70 ± 0.36) × 10-4
γ-ray detector (8.03 ± 0.46) × 10-4 (1.60±0.18) × 10-3 (1.68 ± 0.13) × 10-3 (3.94 ± 0.20) × 10-3 (0.61 ± 0.03) × 10-4
absolute efficiency
(4.33 ± 0.26) × 10-4
at 1634–1636 keV
(2.98 ± 0.18) × 10-4
γ-Detector angle 90° 135° 55° 130° 90°
45°
0°
γ-Detector distance Ranging from 18–20 cm 6.6 cm 9.5 cm 5.55 cm 5.19 cm
to target
135
TABLE 4.24. EXPERIMENTAL CONDITIONS USED IN THE FIVE LABORATORIES MEASURING THE DIFFERENTIAL CROSS-SECTIONS OF THE
23
Na(p,p′γ)23Na: 441 keV & 1636 keV γ-rays AND 23Na(p,α′γ)20Ne: 1634 keV γ-ray (cont.)
Experimental Laboratory
Condition
Target NaF evaporated on Ag 53.7 µg/cm2 of NaCl on NaCl layer (41 µg/cm2) NaF evaporated on Ag NaCl evaporated on Ag
film self-supporting 6.3 µm Mylar + 3.5 nm evaporated on a thin self-supporting thin film. self-supporting thin film
Au layer Ag foil
NaF 34.9 ± 1.5 µg/cm2, Thicknesses of 5.12⋅1017 1.9⋅1017 at/cm2 and

Ag 51.0 ± 1.8 µg/cm2 at/cm2 and 3.21⋅1017 4.8⋅1017 at/cm2 were
at/cm2 were obtained for obtained for Na and Ag
Na and Ag
Beam size 1.0 x 1.0 mm2 4 mm φ Not reported Not reported Not reported
Energy range 2463–4101 keV 1890–3050 keV 1770–3020 keV 2150–4070 keV 1000–2900 keV
Energy step 3–10 keV 15 keV 1–10 keV 2–20 keV not reported
Beam Currents 70–140 nA 10–20 nA 37 nA < 200 nA < 100 nA
Particle detector 150° 165° 135° 157° 165°
angle and solid
3.37 ± 0.08 msr 7.5 ± 0.2 msr 4.3 ± 0.2 msr not reported 0.746 ± 0.007 msr
angle
136
TABLE 4.25. UNCERTAINTY BUDGET FOR THE MEASUREMENTS PERFORMED IN THE
FIVE LABORATORIES MEASURING THE DIFFERENTIAL CROSS-SECTIONS OF THE
23
Na(p,p′γ)23Na: 441 keV & 1636 keV γ-rays AND 23Na(p,α′γ)20Ne: 1634 keV γ-ray
441 keV γ-ray peak area < 2.0% 1.0% < 1.5% 1.4–3.2% 1.4%
1634–1636 keV γ-ray peak area 1–3%, 9–11% b 1–5% 2.0% 1.4–3.2% 3.2%d
p+Ag/Au elastic peak area 1.0% 1.0% < 1.0% < 1.0% 1.0%
p+Ag/Au Rutherford cross-section c 1.0% 1.0% 3% 1.0% 1.0%
Na-to-Ag/Au atomic ratio 2.0% 2.5% 3% 3% 3.7%
Particle detector solid angle 2.4% 2.0% 4.6% 2.5% 1.0%
γ-ray detector absolute efficiency:
at 441 keV 6% 7% 3% 5% 5%
at 1634–1636 keV 6% 11% 8% 5% 5%
23
4.7.1.1. Na(p,pγ1–0)23: 441 keV γ-ray
The excitation function for the 441 keV γ-ray line that was measured by the five laboratories,
is presented in Fig. 4.49. In the energy region below 2500 keV, data from RBI and ATOMKI
are in good agreement. In the energy range from 2500–3000 keV, data from all involved
laboratories can be compared and the difference among the different data sets for similar
angles is less than 20%. In this energy region, the average differential cross-section would be
within an uncertainty range of ± 10%, which is comparable to the uncertainty of the
individual experimental sets. Small energy shifts of 4 keV and 8 keV, which are within the
quoted uncertainties, should be applied to the RBI and LABEC-INFN data, respectively, so
that they match with the other datasets. In the energy range above 2500 keV, NSTRI data are
about 30–40% higher than the other data sets, which are in good agreement with each other.
Above 3000 keV, data have been measured only by LABEC-INFN and CTN/IST and they
agree within 15% except for a small energy region from 4000 to 4050 keV, where the
difference is 25%. Again the average differential cross-section agrees with the individual data
within ± 7.5%, i.e. within the stated experimental uncertainties.
137
FIG. 4.49. Measured differential cross-sections values for the production of 441 keV 23Na(p,pγ1–0)23Na
resonance as a function of proton energy [4.88]. Measurements by five laboratories: a) LABEC-INFN
(45°), b) ATOMKI (55°), c) NSTRI (90°), d) CTN/IST (130°) and e) RBI (135°) .
4.7.1.2. 23
Na(p,pγ2–1)23Na: 1636 γ-ray and 23Na(p,αγ1–0)20Ne: 1634 keV γ-ray
Differential cross-sections for the 1634–1636 keV γ-ray lines from the 23Na(p,αγ1–0)20Ne and
23
Na(p,pγ2–1)23Na reactions were measured by five laboratories and the results are shown in
Fig. 4.50. At energies below 2500 keV, the data from NSTRI are higher by 20–25% then the
data measured from other groups in this energy region. The reason for that could be attributed
to an angular anisotropy of the emitted γ-ray. The differential cross-sections measured at RBI,
CTN/IST and ATOMKI in the energy range from 2500–3000 keV are in good agreement,
since their differences are well within quoted uncertainties. The only exception is the data
from LABEC-INFN, which are systematically lower by 20–25%. In the same energy region,
the data from NSTRI are about 20–25% higher than the data from ATOMKI, CTN/IST and
RBI. As in the case of the 441 γ-ray line, the same shift in energy should be applied to the
data from RBI and LABEC-INFN so that the resonance energies match the data from the
other three groups. For energies above 3000 keV, there exist data from two groups, LABEC-
INFN and CTN/IST, which differ by about 20–25%. The systematic difference observed
between the data from LABEC-INFN and CTN/IST is not the same for the two γ-ray lines,
therefore it is most likely attributable to an error in the determination of the absolute
efficiency of the detector, since this is the only parameter that depends on the γ-ray energy.
138
FIG. 4.50. Measured differential cross-sections values for the production of 1636 and 1634 keV γ-rays
produced by the 23Na(p,pγ2–1)23Na and 23Na(p,αγ1–0)20Ne reactions, respectively, as a function of
proton energy, measured by five laboratories [4.88]: a) LABEC-INFN (45°), b) ATOMKI (55°),
c) NSTRI (90°), d) CTN/IST (130°) and e) RBI (135°).
(b) Data in the literature

Several publications that are dedicated to measurements of proton-induced γ-ray emission
from sodium exist in the literature. These include one measurement of differential cross-
sections [4.89], two different measurements of total cross-sections [4.90, 4.91], and several
measurements of thin [4.92–4.95] and thick target yields [4.1, 4.2, 4.17, 4.96], in the proton
energy range 1–5 MeV, which is of interest for PIGE applications. Recently, evaluated cross-
sections were also published for the 23Na(p,p′γ)23Na reaction, for energies up to 2400 keV
[4.97]. This evaluation, however, only took into account the data that were available in the
literature until 2013. Consequently, not all the data that were measured within this CRP were
included in that evaluation. A new evaluation that includes all the published data from this
CRP is recommended.
Figure 4.51 illustrates the measured differential cross-sections for the production of 441 keV
γ-ray from the 23Na(p,pγ1–0)23Na reaction that were available in the literature and IBANDL,
prior to the CRP [4.89– 4.91]. Since the latter two are given as total cross-sections in the
articles, they have been subsequently divided by a factor 4π to obtain differential cross-
sections. The data from [4.91] are higher than the two other datasets by about a factor of 7 and
are therefore omitted from the subsequent comparison. In Fig. 4.52, the above-mentioned data
from the literature are compared with the data measured in this CRP. As can be seen in the
figure, the data from NSTRI are about 10% higher than the data from Mateus et al. [4.90], in
the low energy region. This difference could be attributed to the angular anisotropy of the
emitted γ-ray, since the two measurements were performed at different angles: 90° by NSTRI
and 120° by [4.90]. At proton energies below 2.5 MeV, the data from ATOMKI and RBI are
in good agreement, however they are about 30% lower than the data from [4.90]. The angles
at which the various measurements have been performed are comparable (55° at ATOMKI,
135° at RBI and 120° by [4.90]). Therefore, the deviations cannot be due only to the
anisotropy in the angular distribution. Data from Caciolli et al. [4.89], which were measured
at an angle of 135°, with quoted overall uncertainty 9%, are systematically lower than the
139
presently measured data, by a relative factor ranging from about 15% (compared to data from
LABEC-INFN) to about 35% (compared to data from ATOMKI). Compared to the older data,
the present CRP measurements were made with smaller energy steps and using thinner
targets, thus revealing a finer structure in the excitation function. The observed fine structure
makes these new data particularly valuable for the extension of the theoretical evaluation of
the 23Na(p,p′γ)23Na cross-section to higher energies.
Figure 4.53 compares the data measured by this CRP with the dataset of Caciolli et al. [4.89]
for the production of 1636 and 1634 keV γ-rays from the 23Na(p,pγ2–1)23Na and 23Na(p,αγ1–
20
0) Ne reactions. The data of [4.89], which has a quoted overall uncertainty of 12% for the
1634–1636 keV γ-ray lines, are systematically lower than the presently measured data. The
underestimation ranges from about 15% (compared to the data from LABEC-INFN) to about
50% (compared to the other data). As already noted above for the differential cross-section of
the 441 keV γ-ray, the large energy steps employed by Caciolli et al. [4.89] lead to excitation
curves that miss the fine structure and sharp resonance features that are clearly observed in the
more detailed new measurements performed during this CRP.
FIG. 4.51. Differential cross-sections for the 1636 keV 23Na(p,pγ2–1)23Na and 1634 keV 23Na(p,αγ1–
20
0) Ne resonances as a function of proton energy. Data taken from the literature a) [4.89] at 135°,
b) [4.90] at 120° and c) [4.91] at 90°. The latter two are given as total cross-sections and must be
divided by a factor 4π to obtain differential cross-sections.
140
FIG. 4.52. Measured differential cross-sections for the 441 keV 23Na(p,pγ1–0)23Na resonance as a
function of proton energy. Data obtained from the laboratories in the CRP [4.88]: a) LABEC-INFN
(45°), b) ATOMKI (55°), c) NSTRI (90°), d) CTN/IST (130°), e) RBI (135°) and from other sources:
f) [4.89] (135°) and g) [4.90] (120°).
FIG. 4.53. Measured differential cross-sections for 1636 keV 23Na(p,pγ2–1)23Na and 1634 keV and
23
Na(p,αγ1–0)20Ne as a function of proton energy. Data obtained from laboratories in the CRP [4.88]:
a) LABEC-INFN (45°), b) ATOMKI (55°), c) NSTRI (90°), d) CTN/IST (130°), e) RBI (135°) and from
other sources: f) [4.89] (135°).
141
23
4.7.2.1. Na(p,pγ1–0)23: 441 keV γ-ray
For the 441 keV γ-ray line, there are five different datasets for thick-target yields available in
the literature [4.1, 4.2, 4.17, 4.88, 4.96]. Chiari et al. [4.88] reported thick-target yields for
three different laboratory angles (0º, 45º and 90º) and showed that the difference in the thick-
target yield that exists for these three angles is due to the anisotropic emission of 441 keV γ-
ray line. Thick-target yields were also calculated from the new experimental differential
cross-section data using the same computer program distributed to CRP participants by [4.34].
This program computes the number of γ-rays produced by a certain reaction, assuming a
homogeneous thick target and taking into account the beam energy loss and the differential
cross-section of the reaction. Figure 4.54 shows thick-target yields calculated using the
experimental data measured in the present CRP and the available thick-target yield data [4.1,
4.2, 4.17, 4.88, 4.96]. In order to compare differential cross-sections with the thick-target
yield datasets, the integrated cross-section data were normalized to the data of Kiss et al. [4.2]
at the energy of 2.4 MeV. The normalization was performed by adding a constant number to
the integral to take into account that the measured differential cross-sections do not extend
down to zero energy. However, it can be seen that some sets of data can’t be normalized to
the data of Kiss et al. [4.2] at the energy of 2.4 MeV without subtracting some value, which is
physically not possible.
FIG. 4.54. Comparison of the thick-target yields for 441 keV 23Na(p,p′γ)23Na as a function of energy.
Calculated yields taken from measured cross-sections from laboratories in the CRP [4.88]: (a) RBI
(135°), (b) CTN/IST (130°), (c) NSTRI (90°), (d) LABEC-INFN, (e) ATOMKI (55°), and measured
yields taken from (f) [4.96] (90°), (g) [4.17] (90°), (h) [4.88] (90º),(i) [4.2] (55°) and (j) [4.1] (55°).
Figure 4.55 shows thick-target yields calculated using the experimental data measured in the
present CRP, thick-target yields calculated using previously measured experimental data
[4.17, 4.89–4.91, 4.96] and the available thick-target yield data [4.1, 4.2, 4.17, 4.88, 4.96].
142
FIG. 4.55. Comparison of the thick target yields for 441 keV 23Na(p,p′γ)23Na as a function of incident
energy. Calculated yields are from measured cross-sections of laboratories in the CRP [4.88]:
(b) RBI (135°), (c) [4.90] (120°), (d) CTN/IST (130°), (e) NSTRI (90°), (g) LABEC-INFN (90°),
(h) ATOMKI (55°) and from other sources: (a) [4.89] (135°), (f) [4.91] (90°) and (i) SigmaCalc.
Measured yields are from (j) [4.96] (90°), (k) [4.17] (90°), (l) [4.88] (90º),(m) [4.2] (55°) and (n)
[4.1] (55°).

It should be noted that 23Na is one of the cases where the data for the γ-ray production cross-
sections are in good agreement, and have relatively low uncertainties. All the new
measurements agree within the quoted uncertainties in the measured energy range and
therefore can be used for analytical purposes. Since slight anisotropy was observed, it is
advisable to use the cross-sections at the appropriate angle.
4.7.2.2. 23
Na(p,pγ2–1)23Na: 1636 γ-ray and 23Na(p,αγ1–0)20Ne: 1634 keV γ-ray
For the 1636 and 1634 keV γ-rays lines, there is only one dataset for thick-target yields
available in the literature. Chiari et al. [4.88] reported thick-target yields for three different
laboratory angles: 0º, 45º and 90º. Differences in the thick-target yield are due to the
anisotropic emission of the γ-ray line. Figure 4.56 shows thick-target yields calculated using
the experimental data measured in the present CRP, thick-target yields calculated using
previously measured experimental cross-sections [4.89] and the available measured thick-
target yield data [4.88] for 90º.
143
FIG. 4.56. Comparison of the thick target yields for the 1636 keV 23Na(p,pγ2–1)23Na and 1634 keV
23
Na(p,αγ1–0)20Ne γ-rays as a function of energy. Calculated yields from cross-section data from
laboratories in the CRP [4.88]: b) LABEC-INFN (90°), c) ATOMKI (55°), d) CTN/IST (130°),
e) NSTRI (90°), f) RBI (135°) and other sources a) [4.89] (135°). Measured yields from g) [4.88]
(90°).

All the new measurements agree within the quoted uncertainties in the low energy range (up
to 3 MeV) and therefore can be used for analytical purposes in this energy region. However,
care should be taken when analysis is performed at energies higher than 3 MeV since
descrepancies are observed between existing datasets.
It is recommended that data for both γ-lines are evaluated over the entire measured energy
range as well as for the γ-line 1636 + 1634. In addition, it is recommended that thick-target
yield measurements be performed to resolve existing differences.
4.8. MAGNESIUM
4.8.1. Gamma-ray production cross-sections of magnesium from proton-induced

nuclear reactions
For the analysis of Na, Mg and Al elements, which are common components of
archaeological and cultural heritage materials such as ceramic, glass, obsidian and paper, it is
recommended that PIGE analysis be performed simultaniously with PIXE [4.98, 4.99], since
the two are considered to be complementary techniques. However, γ-ray producing cross-
section data, which is required for PIGE analysis of Mg, are rarely found in the literature.
Therefore, measurements of these cross-sections were considered a priority for the CRP.
144
A review of the literature that existed prior to the CRP revealed the following publications on
Mg: “A study of relative thick-target yield of prompt γ-rays at some specific proton energies”
by Anttila et al. [4.1], in which the most suitable γ-rays for elemental analysis have been
suggested. Excitation function measurements of thick-target γ-ray yield for Mg have been
presented by Savidou et al. [4.17] in the proton energy interval 1.0−4.1 MeV at θlab = 90°.
Lastly, Boni et al. [4.5] have measured differential cross-sections of 390, 585, 975, 1612,
1809 and 1369 keV γ-rays from Mg at θlab = 90° at 2.2-3.8 MeV proton energies.

Four laboratories were involved in measurements of γ-ray producing cross-sections from the
reactions 24Mg(p,p′γ)24Mg (Eγ = 1369 keV γ-ray) and 25Mg(p,p′γ)25Mg (Eγ = 390, 585 and 975
keV γ-rays): CTN/IST, RBI, NCSR “Demokritos”, and NSTRI. One laboratory, NSTRI
(Iran), performed measurements of the 1014 keV γ-ray from the 26Mg(p,γ2–0)27Al reaction.
In a typical measurement of differential cross-sections using Mg targets, the sources of the

systematic uncertainties are: the γ-ray detector absolute efficiency (due to the uncertainties of
calibration for radioactive sources and fitting the experimental data), the stoichiometric ratio,
particle detector solid angle. The total systematic uncertainty was obtained by taking the
quadratic sum of all these uncertainties. The uncertainties in the proton scattering cross-
section values of Ag can be considered negligible, since these cross-sections are purely
Rutherford. Different methods can be used to prepare a thin target of Mg if the traget is not
purchased from a commercial company, which may lead to different uncertainties in the
quoted target thickness. The laboratories participating in the CRP adopted different
experimental conditions, which are summarized in Table 4.26.
TABLE 4.26. EXPERIMENTAL CONDITIONS OF THE LABORATORIES INVOLVED IN THE

DIFFERENTIAL GAMMA-RAY PRODUCTION CROSS-SECTION MEASUREMENTS FOR Mg
Reference Data Angle Ep Target Quoted Data
source (MeV) uncertainties presentation
[4.78] IBANDL 130° 0.86– Mg thin film 7% Tabular (R33
4.02 evaporated on file)
systematic
Ag thin film
[4.39] IBANDL 55°, 2.41– — statistical, Tabular (R33
90° 4.55 systematic file)
[4.100] Author 90° 1.98–3.0 Mg (60 µg/cm2) 5% statistical, Tabular (R33
evaporated on 6% systematic file)
Ag (40 µg/cm2)
[4.79] IBANDL 135° 1.93– Au (4nm) Overal Tabular (R33
3.05 deposited on uncertainty 8– file)
MgF2 (54.1 42%
µg/cm2)
deposited on
(6.3 µm) Mylar

foil
145
24
4.8.1.1. Mg(p,pγ1–0)24Mg: 1369 keV γ-ray
The 1369 keV γ-ray is emitted from the first excited state of 24Mg, which has a 2+ spin.
Therefore the angular distribution of this γ-ray is anisotropic. The measured PIGE differential
cross-sections of the 1369 keV γ-ray from the 24Mg(p,p′γ)24Mg reaction are shown in Fig.
4.57. As illustrated in the figure, the PIGE cross-sections in the 1.0–3.0 MeV energy range are
low and nearly smooth, except around energies 2409, 2919 and 2979 keV, where strong
resonances are observed. The 2919 keV and 2979 keV resonances have also been reported by
Boni et al. [4.5] and all three resonances have been observed by Zamboni et al. [4.79].
Furthermore, a weak resonance at 2009 keV has been reported by [4.79], although the results
could not be validated by [4.100], since the selected energies were not coincident with this
specific energy.
20 a
b
15
dσ/dΩ (mb/sr)
10
0
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200
Energy (keV)
FIG. 4.57. Differential γ-ray production cross-sections of 1369 keV 24Mg(p,p′γ)24Mg as a function of
proton energy. Data obtained from a) [4.100] (90°) and b) [4.79] (135°).
As can be seen in Fig. 4.57, the measured data are in close agreement in both the energy
position of the resonances and the cross-section values with the data of Zamboni [4.79]. As
mentioned in [4.79], the cross-section data values in [4.5] should be multiplied by a factor of
2π in order to compare them to the measured data.
25
4.8.1.2. Mg(p,p'γ)25Mg: 390, 585, 975 keV γ-rays
Cross-sections for Eγ = 390, 585 and 975 keV from the 25Mg(p,p′γ)25Mg reaction show narrow
resonances at proton energies higher than 1850 keV. In Figs. 4.58–5.60, the data of
Shirifzadeh et al. [4.100] are compared with those reported by Zamboni et al. [4.79] and
Preketes-Sigalas et al. [4.39]. In the case of the 585 keV γ-ray line, the comparison also
includes the data of Jesus et al [4.78].
146
12
(a)
a
10
b
8 c
dσ/dΩ (mb/Sr)
6
(b)
4
0
800 1800 2800 3800 4800
(c
Energy(keV) )
proton energy. Data obtained from a) [4.100] (90°) b) [4.79] (135°), c) [4.39] (90°) and d) [4.39]
(55°).
14
a
12
b
10
c
8
dσ/dΩ (mb/Sr)
d
6
0
800 1800 2800 3800 4800
Energy(keV)
proton energy. Data obtained from a) [4.100] (90°) b) [4.79] (135°), c) [4.39] (90°), d) [4.39] (55°)
and e) [4.78] (130°).
147
12
10
dσ/dΩ (mb/Sr)
8
0
800 1800 2800 3800 4800
Energy(keV)
proton energy. Data obtained from a) [4.100] (90°), b) [4.79] (135°), c) [4.39] (90°) and d) [4.39]
(55°).
(b) Assessment
In Fig. 4.58, at energies higher than 2.24 MeV, the excitation function is dominated by
resonances. The data of Shirifzadeh et al. [4.100] reveal the presence of two resonances for
390, 585 and 975 keV γ-rays from the 25Mg(p,p′γ)25Mg reaction at 2159 keV and 2789 keV as
well as two more resonances for 585 keV γ-rays observed at 2319 and 2569 keV, which are
missing in the data of [4.79], due to their larger energy steps.
Table 4.27 compares the measured energy resonances of 25Mg(p,p'γ)25Mg reaction in the data
of [4.100] with those published in [4.5] and [4.79]. As mentioned in [4.17], the 585 keV γ-line
is a suitable choice for the elemental Mg analysis over the whole energy range, due to its
relative high cross-section values compared to the other γ-lines. Other reasons include the low
background underneath this peak and the absence of overlapping reaction channels from other
light elements. Moreover, the 2.60−2.65 keV proton energy interval for the analysis of thin
multi-element samples containing Mg is recommended. This energy range is greater than the
energy loss of protons in thin samples and furthermore the PIGE cross-section values are high
and constant enough within this range.
148
25
TABLE 4.27. COMPARISON OF THE ENERGY RESONANCES OF THE Mg(p,p′γ)25Mg
REACTION: PUBLISHED IN THE LITERATURE VS MEASURED
Published Ep [4.5, 4.79]
Measured Ep (keV)
390 keV 585 keV 975 keV
2039 2041 2060 2041
2159 — — —
2279 2259 2275 2259
a
2319 — — —
2349 2351 2351 2352
2399 2399 2399 2399
2509 2507 2514 2507
2569a — — —
2589 2589 2589 2589
2689 2678 2678 2678
2739 2740 2740 2740
2789 — — —
2857 2857 2857
2864
2858 ± 5 b 2856 ± 5 b 2855 ± 5 b
2901 2901 2901
2899
2900 ± 5 b 2902 ± 5 b 2897 ± 5 b
2962 2947 —
2959
2960 ± 5 b 2956 ± 5 b
a
only visible for the 585 keV γ-line
b
resonance energies from Ref. [4.5]
26
4.8.1.3. Mg(p,γ2–0)27Al: 1014 keV γ-ray
The measured PIGE differential cross-sections of the 1014 keV γ-rays from the
26
Mg(p,γ2–0)27Al reaction are shown in Fig. 4.61 [4.100]. To the best of our knowledge, there
was no previous cross-section data in the literature for the 1014 keV γ-line. The PIGE cross-
sections for this γ-ray line are low and smooth over almost the entire energy range, except at
Ep = 2879 keV where a strong resonance is observed.
149
30
25
dσ/dΩ(mb/sr)
20
15
10
0
1700 1900 2100 2300 2500 2700 2900
Energy (keV)
26
FIG. 4.61. Differential γ-ray production cross-sections of 1014 keV Mg(p,γ2–0)27Al measured at an
angle of 90° as a function of proton energy [4.100].
The differential cross-section data shown in Figs. 4.58–4.60 were used to calculate thick-
target yields using the computer program of [4.34]. In the calculation of thick-target yields, if
the energy in the integration starts from a certain energy above zero that corresponds to a
given reaction cross-section value, the measured thick-target yields up to that certain energy,
have to be added to the integral. For this purpose, the data sets reported by Kiss et al. [4.2]
were employed. The resulting thick-target yields are compared in Figs. 4.62–4.66.
150
4.5E+06
a
4.0E+06
b
dσ/dΩ(mb/sr)
3.5E+06
c
3.0E+06
d
2.5E+06
e
2.0E+06
1.5E+06
1.0E+06
5.0E+05
0.0E+00
0 1000 2000 3000 4000
Energy (keV)
FIG. 4.62. Comparison of measured and calculated thick-target yields for the 1369 keV γ-ray
produced by the 24Mg(p,p′γ)24Mg reaction as a function of energy. Calculated yields obtained from: a)
[4.100] (90°), b) [4.79] (135°), c) [4.5] (90°), while measured yields are taken from: d) [4.2] (55°)
and e) [4.1] (55°).
8.E+05
a
7.E+05
b
6.E+05 c
dσ/dΩ(mb/sr)
5.E+05 d
4.E+05
3.E+05
2.E+05
1.E+05
0.E+00
0 1000 2000 3000 4000 5000
Energy (keV)
FIG. 4.63. Comparison of measured and calculated thick-target yields for the 390 keV γ-ray produced
by the 25Mg(p,p′γ)25Mg reaction as a function of energy. Calculated yields obtained from: a) [4.100]
(90°), b) [4.79] (135°), c) [4.39] (55°) and d) [4.5] (90°).
151
2.0E+06
a
1.8E+06
1.6E+06 b
1.4E+06 c
dσ/dΩ(mb/sr)
1.2E+06 d
1.0E+06 e
8.0E+05 f
6.0E+05
g
4.0E+05
2.0E+05
0.0E+00
0 1000 2000 3000 4000 5000
Energy (keV)
by the 25Mg(p,p′γ)25Mg reaction as a function of energy. Calculated yields obtained from: a) [4.100],
b) [4.79], c) [4.78], d) 55̊ [4.39], e) 90̊ [4.39], f) [4.5] and measured yields taken from g) [4.17].
9.0E+05
a
8.0E+05
b
7.0E+05
c
dσ/dΩ(mb/sr)
6.0E+05
5.0E+05 d
4.0E+05 e
3.0E+05
2.0E+05
1.0E+05
0.0E+00
0 1000 2000 3000 4000 5000
Energy (keV)
by the 25Mg(p,p′γ)25Mg reaction as a function of energy. Calculated yields obtained from: a) [4.100]
(90°), b) [4.79] (135°, c) [4.39] (55°), d) [4.39] (90°) and e) [4.5] (90°).
152
1.4E+06
a
1.2E+06
dσ/dΩ(mb/sr) b
1.0E+06 c
8.0E+05 d
6.0E+05
4.0E+05
2.0E+05
0.0E+00
0 1000 2000 3000 4000 5000
Energy (keV)
FIG. 4.66. Comparison of measured and calculated thick-target yields for the 1014 keV γ-ray
produced by the 26Mg(p,γ2–0)26Mg reaction. Calculated yields taken from a) [4.100] (90°), b) [4.5],
while measured yields from c) [4.2] (55°), and d) [4.1] (55°).

From Fig. 4.62, it is clear that the calculated thick-target yields for the 1369 keV - ray from
24
Mg(p,p′γ)24Mg reaction are neither in agreement with each other, nor with the yields derived
from the data reported by Boni et al [4.5]. The reason is that there are some narrow
resonances around 2010 keV and 2410 keV, but the cross-sections are not measured with
small enough energy steps to be sensitive to them. As a result, none of these measured
differential cross-sections are recommended for PIGE analysis.
According to the results shown in Figs. 4.63–4.65 for the 390, 585 and 975 keV γ-rays
produced from the 25Mg(p,p′γ)25Mg reaction, the data of Shafifzadeh et al. [4.100] is
recommended for proton energies less than 3 MeV, due to good agreement with the thick-
target yield measurements over an extended energy range. For energies above 2.5 MeV, the
measured data by K. Preketes-Sigalas et al. [4.39] give the best agreement with the thick-
target yields and therefore they are recommended. Furthermore, in the energy region from 2.0
to 3.0 MeV, the data measured by Zamboni et al. [4.79] are in good agrement with the
measured data of Savidou et al. [4.17]. However, for the 585 keV γ-ray line, the data reported
by Jesus et al. [4.78] seem to overestimate the thick-target yields that are calculated and those
that are measured by Savidou et al. [4.17].
For the 1014 keV γ-ray produced by the 26Mg(p,γ2–0)26Mg reaction, there is no reasonable
overlap between the energy ranges of the various thick-target yield measurements and the
energy range of the differential cross-sections measurements by [4.100]. Therefore no
comparison can be performed and hence none of these data are recommended for PIGE
analysis.
153
4.9. ALUMINUM
4.9.1. Gamma-ray production cross-sections of aluminum from proton-induced nuclear

reactions
Gamma-rays of 844 and 1014 keV are produced following the inelastic scattering of protons
by aluminum. The cross-sections of these processes are large, which leads to high γ-ray yields
that may be used to quantify aluminum concentrations in a large variety of samples when
using proton beams. Other γ-rays (of higher energy) are also produced during the
bombardment, by the process called proton capture. However, since this process has a much
smaller cross-section than the inelastic scattering process and the efficiency of the γ-ray
detector decreases with increasing γ-ray energy, inelastic scattering gives a much higher
sensitivity for analysis of in-depth homogeneous samples. In contrast, proton capture excites
several very narrow and rather isolated resonances, which are frequently used for the purpose
of depth analysis. Another nuclear reaction that may occur is the (p,α) reaction, but the γ-ray
produced by this reaction may also be obtained by bombardment of magnesium by protons.
Therefore, it is not suitable for analytical purposes, unless it can be ensured that magnesium is
not present in the sample. Furthermore, 26Mg(p,γ)27Al may also produce the 844 and 1014
keV γ-ray lines, but since this reaction is less likely to occur, it’s contribution will be
significant only for samples very rich in magnesium. The amount of magnesium may be
controlled by the 25Mg(p,p′γ)25Mg reaction, which produces a different γ-ray line.
Regarding the availability of cross-section data, inelastic scattering cross-sections and proton
capture cross-sections are scarce. Although this reaction has been thoroughly studied, most of
the studies have focused on measuring resonance strengths. As a result, considerable effort
was made during this CRP to measure γ-ray producing cross-sections and thick-target yields
for the inelastic scattering of protons by aluminum.
27
4.9.1.1. Al(p,pγ1–0)27Al: 844 keV γ-ray
Some of the results of these measurements have been described in the section on the
interlaboratory exercise (Section 3.6). In this section, more details of the different
experimental set-ups, the complete datasets as well as a comparison with datasets found in the
literature are presented.

Measurements of the 27 Al(p,p′γ1–0)27Al reaction were performed in the following laboratories:
CTN/IST, LABEC-INFN, RBI and NSTRI. The experimental conditions for the different
laboratories are presented below.
CTN/IST
The proton beams were accelerated by a 2.5 MV Van der Graaff accelerator and a 3 MV
Tandetron accelerator. The 2.5 MV Van der Graaff accelerator delivered proton beams with
energies in the range 0.30 to 2.40 MeV. Proton beam currents of 250 nA were employed. The
proton energy was calibrated by the 668.0, 872.1 and 1373.2 keV resonances produced by the
19
F(p,αγ)16O reaction and also by the 1645.1 and 1930.7 keV resonances produced by the
23
Na(p,p′γ)23Na reaction. Proton beams, produced by a Duoplasmatron source, were
accelerated by the 3 MV Tandem accelerator, with energies in the range of 2.0 to 3.0 MeV.
Proton beam currents of the order of 100 nA were employed. The proton energy was
calibrated by the 1645.1 and 1930.7 resonances produced by the 23Na(p,p′γ)23Na reaction and
by the 3470 keV resonance produced by the 16O(p,p)16O reaction. The same reaction chamber
154
was used for the two sets of measurements. The reaction chamber is electrically insulated
from the beam line and works together with the target holder and beam stopper as one
Faraday cup for beam charge collection. Scattered protons at 157° to the incident beam were
detected by Camberra PIPs detectors with active area of 50 mm2 and depletion layer 100 mm,
located in Cornell geometry at 88 mm from the target, with solid angles defined by apertures
of 6.00 ± 0.05 mm. Gamma-rays were detected by a 45% Ge(HP) detector located at 130° in
relation to the incident beam direction and at a distance of 55.5 mm from the target. The
absolute efficiency versus γ-ray energy was obtained by measuring the yields of γ-rays
produced by ion sources (133Ba and 152Eu) that were calibrated in activity as well as by Monte
Carlo simulations using GEANT and PENELOPE. The target was produced by evaporating a
thin film of pure Al onto a thin self-supporting Ag film. RBS performed on the Ag film was
used to normalize the cross-sections. The uncertainty budget can be found in Table 4.28.
LABEC – INFN
The measurement was performed with a 3 MV Tandetron accelerator. Proton beam currents
of 100 nA were employed. The proton energy was calibrated by the resonances at the 992
keV γ-ray produced by the 27 Al(p,γ) reaction, the 1683 keV γ-ray produced by the 27 Al(p,p′γ)
reaction, the 2876 keV γ-ray produced by the 27Al(p,p) reaction and the 3470 keV resonance
produced by the 16O(p,p)16O reaction. Gamma-rays were detected by a 50% Ge(HP) detector
located at 90° in relation to the incident beam direction and at a distance of 205 mm from the
target. The absolute efficiency versus γ-ray energy was obtained by measuring the yields of γ-
rays produced by ion sources (152Eu). Elasticly scattered protons were collected
simultaneously using a Si-PIN diode detector with area of 10 mm2, thickness of 300 mm and
energy resolution 13 keV FWHM, placed at a backward angle of 150° and collimated by a 1 ×
13 mm2 vertical slit. The subtended solid angle was 3.37 ± 0.08 msr. The target was produced
by evaporating a thin film of pure Al onto a thin self-supporting Ag film. RBS by the Ag film
was used to normalize the cross-sections. The uncertainty budget can be found in Table 4.28.
RBI
The protons were accelerated by a 6.0 MV Tandem Van der Graaff. The energy was
calibrated by the 992 keV resonance produced by the 27Al(p,γ) reaction and the 7Li(p,n)7Be
threshold at Ep = 1880.7 ± 0.4 keV. Gamma-rays were detected by a 20% Ge(HP) detector
located at 135° in relation to the incident beam direction and at a distance of 100 mm from the
rays produced by ion sources (60Co, 133Ba, 137Cs and 152Eu). Elastically scattered protons
were collected simultaneously using a silicon surface barrier detector placed at the backward
scattering angle of 165°, with a solid angle of 7.5 ± 0.2 msr. The target was a thin film of pure
Al evaporated onto a thin self-supporting Mylar film, with a 4 nm Au layer. RBS by the Au
layer was used to normalize the cross-sections. The uncertainty budget can be found in Table
4.28.
NSTRI
The protons were accelerated by a 3.0 MV Van der Graaff accelerator. The proton energy was
calibrated by the 992 keV resonance produced by the 27Al(p,γ) reaction and the 7Li(p,n)7Be
threshold at Ep = 1880.7 ± 0.4 keV. Gamma-rays were detected by a 50% Ge(HP) detector
located at 90° in relation to the incident beam direction and at a distance of 51.9 mm from the
rays produced by ion sources (60Co, 133Ba, 137Cs, 152Eu and 241Am). Elastically scattered
protons were collected simultaneously using a silicon surface barrier detector placed at the
backward scattering angle of 165°, with a solid angle of 0.746 msr. The target was created by
155
evaporating a thin film of pure Al onto a thin self-supporting Ag film. RBS by the Ag film
was used to normalize the cross-sections. The uncertainty budget may be found in Table 4.28.
Since every laboratory used the same method to measure the cross-sections, the uncertainty
budget is the same for all the laboratories and is listed in Table 4.28.
TABLE 4.28. UNCERTAINTY BUDGET FOR THE MEASUREMENTS OF THE CROSS-

SECTIONS OF Al MADE IN ALL LABS
γ-ray peak area 1–3%
p+Ag elastic peak area 1%
p+Ag Rutherford cross-section 1%
Al/Ag atomic ratio 3%
Particle detector solid angle 2.5%
γ-ray detector efficiency 6%
Proton beam energy 0.1%
Overall uncertainty 8–10%
Figure 4.67 shows a comparison of cross-section values obtained by the above listed groups.
These values may also be found in the IBANDL database: Jesus et al. (CTN/IST) [4.78] from
1497 to 3000 keV in steps from 1 keV (on resonance) to 10 keV (off resonance); Chiari et al.
(LABEC-INFN) [4.101] from 2507 to 4089 keV in steps from 5 keV (on resonance) to 10
keV (off resonance); Zamboni et al. (RBI) [4.79] from 1932 to 2507 keV in 16 keV steps and
from 2507 to 3053 keV in steps from 5 keV (on resonance) to 15 keV (off resonance); Jokar
et al. (NSTRI) [4.102] from 1497 to 2388 keV in 20 keV steps and from 2388 to 2998 keV in
steps from 5 keV (on resonance) to 10 keV (off resonance). Figure 4.67 pertains to an energy
region covered by the four measurements. The energy region from 2.5 to 3.0 MeV has been
covered by all four measurements and, as is illustrated in Fig. 4.67, these measurements agree
within the experimental uncertainty for both on and off resonance. Around 2.5 MeV, the
proton energy scales seem to be correct for all the measurements except Jokar et al., which
are 3 keV lower. However around 2.9 MeV, Jokar et al. and Jesus et al. present energies 3
keV higher than Chiari et al. and 6 KeV higher than Zamboni et al. At an energy range lower
than 2.5 MeV, as illustrated in Fig. 4.68, the energy steps of Zamboni et al. and Jokar et al.
are too large to pick up most of the resonances. However, with respect to some gross features
of the excitation function, the results of Zamboni et al. and Jokar et al. agree with Jesus et al.
within the experimental uncertainty.
156
FIG. 4.67. Comparison of cross-sections for the 844 keV γ-ray produced by the 27Al(p,pγ1–0)27Al
reaction, as a function of proton energy in the range of 2250 – 3050 keV. Data obtained within the
CRP: a) [4.78], b) [4.101], c) [4.102] and d) [4.79]
FIG. 4.68. Same as Fig. 4.67 but for the proton energy range 1490–2250 keV.
Only Chiari et al. [4.101] have made measurements for the energy range above 3050 keV.

A review of the existing literature produced one dataset measured by Boni et al. [4.5] (values
may be found in the IBANDL database). Cross-sections were measured in the proton energy
range from 2190 to 3800 keV in steps of around 27 keV. In the energy range from 2190 to
3050 keV, Boni et al. missed most of the resonances due to the very large energy step they
157
employed. The cross-section values between the resonances, however, agree with the new
measured values within the experimental uncertainty. Their results are compared with the data
by Chiari et al. [4.101], in the energy range from 3050 to 4100 keV, in Fig. 4.69. As can be
seen in the figure, the data agree within the experimental uncertainties. At 3050 keV, there is
a discrepancy larger than 10 keV that decreases with increasing energy.
reaction, as a function of proton energy in the proton energy range 3050 – 4050 keV. Data obtained
within the CRP a) [4.101] and previous results: b) [4.5].

From the comparison of the results from new and previous measurements, one may conclude
that there is an overall agreement among the data within the experimental uncertainty.
Therefore, it is recommended to use a combined excitation function in the energy range from
1490 to 4100 KeV, composed of the data of Jesus et al. [4.78] from 1490 to 2500 keV and the
data of Chiari et al. [4.101] from 2500 to 4100 keV, in PIGE bulk analysis.
27
4.9.1.2. Al(p,pγ2–0)27Al: 1014 keV γ-ray

Measurements of the 27Al(p,pγ2–0)27Al reaction were performed simultaneously with the 844
keV γ-ray at the same laboratories, with the same conditions and uncertainty budget
mentioned above. The energy range and steps are also the same. Additionally, results from
Kiss et al. (ATOMKI) [4.78] were also obtained. The measurement was performed with a 5
MV Van de Graaff accelerator. Proton beam currents of 300 nA were employed. The proton
energy was calibrated using the 992 keV γ-ray produced by the 27Al(p,γ) reaction. Gamma-
rays were detected by a 40% Canberra GR4025–7600SL detector located at 55° in relation to
the incident beam direction and at a distance of 95 mm from the target. The absolute
efficiency versus γ-ray energy was obtained by measuring the yields of γ-rays produced by
ion sources (152Eu and 60Co). Elastically scattered protons were collected simultaneously
using an ion implanted Si detector with active depth 500 µm and energy resolution 13 keV,
158
placed at an angle of 135° relative to the incident beam direction and at a distance of 4.3 ± 0.1
cm from the target. A copper collimator with a circular aperture of 3.2 mm diameter was
mounted in front of the Si detector. The solid angle of the Si detector was 4.3 ± 0.2 msr, from
geometrical calculations. The target was produced by evaporating a thin film of pure Al onto
a thin self-supporting C film. The absolute values of the charge and of the areal density of Al
atoms were used to calculate the cross-sections.
Although the 844 keV γ-line, which originates from a spin 1/2 state, is expected to be
isotropic, the same does not apply to the 1410 keV γ-line, which originates from a spin 3/2
state and which may display anisotropy. Two of the measurements were set at an angle of 90°
and the other two measurements at angles of 130° and 135°. Since the latter two angles are
very similar, the measurements are considered to be performed under the same conditions.
Chiari et al. also measured at 45°, and the results were shown to agree, within the
experimental uncertainty, with the data obtained at 90°.
Figures 4.70 and 4.71 show a comparison of the results obtained within this CRP. As can be
seen in Fig. 4.70, in the energy range from 2.5 to 3.0 MeV, for both on and off resonance
regions, the measurements agree within the experimental uncertainty, although they
correspond to different angles. However, it is not clear whether there is a discrepancy in the
energy scale, since some of the measurements seem to have missed the maxima of the
resonances.
In the energy range below 2.5 MeV, the energy steps of Zamboni et al. [4.79] and Jokar et al.
[4.102] were too large. Consequently, their results miss some of the resonances as can be seen
in Fig. 4.71. Overall, however, the gross features of these excitation functions agree with the
results of Jesus et al. [4.78], within the experimental uncertainty.
reaction, as a function of proton energy in the range 2500 – 3050 keV. Data obtained within the CRP:
a) [4.78] (130°), b) [4.101] (45°), c) [4.102] (90°), d) [4.79] (135°), and e) [4.78] (55°).
159
reaction, as a function of proton energy in the range 1640 – 2500 keV. Data obtained within the CRP
by: a) [4.78] (130°), b) [4.102] (90°), and c) [4.79] (135°).

A review of the existing literature on similar measurements resulted in one published set of
measurements [4.5] (data available in IBANDL). The measurements were performed in the
proton energy range from 2190 to 3800 keV in steps of around 27 keV, with a detection angle
at 90°. In the energy range 2190 to 3050 keV, the data of Boni et al. miss most of the
resonances due to the very large energy step employed, but between resonances, the cross-
sections agree with the new measured cross-sections, within the experimental uncertainty. At
the energy range from 3000 to 4100 keV, the Boni data [4.5] are compared with the data of
Chiari et al. [4.101] (Fig. 4.72). Overall, there is good agreement within the experimental
uncertainty, although there is a discrepancy of more than 10 keV in the energy scale at around
3050 keV, which decreases with increasing energy.
160
FIG. 4.72. Same as Fig. 4.71, but for the range of 3000 – 4100 keV. Data obtained within the CRP
a) [4.101] (45°) and previous results b) [4.5] (90°).

From the above comparison of results from new measurements and previous measurements, it
can be concluded that there is an overall agreement within the experimental uncertainty,
which is independent of the detection angles and implies isotropy at the covered energy range
for the 1014 keV γ-line.
27
4.9.1.3. Al(p,αγ1–0)24Mg: 1369 keV γ-ray

Although not very useful for analysis, the 1369 keV γ-line produced from the 27 Al(p,aγ1–
24
0) Mg reaction is obtained in the same spectra as the 844 keV and 1014 keV γ-ray lines
produced from the 27Al(p,p′γ)27Al reaction, as previously mentioned. However, the 1369 keV
γ-line has only been quantified by Jokar et al. [4.102], Zamboni et al. [4.79], and Cedreki et
al. [4.78]. A comparison of the different results is shown in Fig. 4.73. As can be seen, in the
energy range from 1.9 to 2.65 MeV, for both on and off resonance regions, the measurements
agree within the experimental uncertainty, even if they correspond to different detection
angles. At higher energies, however, there are discrepancies in the proton energy scale among
the different measurements. In addition, since the energy step that was used is larger than the
width of some resonances, the measurements have missed the resonances. Therefore, it is not
possible to compare the data in any meaningful way. Although this γ-ray is anisotropic based
on the spin of the initial excited state of 24Mg being 2+, the effect of anisotropy (which may
differ from resonance to resonance) is not expected to be significant, since the thick-target
yields obtained by Chiari et al. [4.101] at different angles indicate very small anisotropies (see
following sections).
161
FIG. 4.73. Comparison of cross-sections for the 1369 keV γ-ray from the 27Al(p,αγ)24Mg reaction as a
function of the proton energy in the range of 1850 – 3050 keV. Data obtained within the CRP by
a) [4.79] (135°), b) [4.102] (90°) and c) [4.78] (55°).

A review of the existing literature revealed two different measurements, those of Boni et al.
[4.5] and Warsh et al. [4.104] (available in IBANDL), which covered the proton energy range
from 2200 to 3800 keV and 3474 to 12152 keV, respectively. In both measurements, an
energy step between 20 to 30 keV was used and the detection angle was 90°. In Fig. 4.74, one
can see the comparison of new and previous measurements for the energy range from 2150 to
3800 keV. Boni et al. [4.5] miss some of the resonances due to the very large step employed,
but between resonances, the different data sets agree within the experimental uncertainty. In
the energy range from 3000 to 3800 keV, the results are compared with the data of Warsh et
al., which reveals a large discrepancy in the energy scale that does not allow for a meaningful
comparison of the two measurements. However, the cross-sections of Warsh et al. [4.104]
seem to be higher than the ones of Boni et al. [4.5].
162
FIG. 4.74 Comparison of cross-sections for the 1369 keV γ-ray from the 27Al(p,αγ)24Mg reaction as a
function of proton energy in the range of 2150 – 3800 keV. Data obtained within the CRP are, from a
to c, the same as for Fig. 4.73, while previous results are obtained from: d) [4.5] (90°) and e) [4.104]
(90°).
(c) Conclusions
From the results above, it can be concluded that it is essential to use an adequate energy step
and appropriate target thickness, in order to obtain a detailed excitation function of a nuclear
reaction. As it stands, it is not possible to employ the results shown above for new and old
measurements for PIGE analysis.
(d) Other γ-ray lines

Although the 27 Al(p,γ)28Si reaction produces several γ-rays that have been thoroughly studied,
most of the published results refer to resonance strengths. There is one exception: the work by
Harissopoulos et al. [4.105] (available in IBANDL), which provides γ-ray yields
corresponding to a combined γ-ray measurement. In order to convert these yields to an
individual γ-ray producing cross-section at a particular angle, one must use branching ratios
and angular distributions. For analysis of in-depth homogeneous samples, there is no
advantage in using this reaction since the cross-sections are much lower than the inelastic
scattering cross-sections. Furthermore, the lowest energy γ-ray line that can be extracted is
1779 keV, which is more energetic than the other widely used (p,p′γ) γ-ray lines and therefore
has a lower detection efficiency. However, the (p,γ) cross-section has very narrow resonances
that have often been used for depth analysis or profiling, e.g. the resonance at 991.92 keV,
which is used as a reference resonance for accelerator calibration.
163
27
4.9.2.1. Al(p,pγ1-0)27Al: 844 keV γ-ray

Thick-target yields were also obtained during this CRP, at the same laboratories and using the
same experimental conditions referred in previous sections. Thick-target yield results depend
on the absolute value of the collected charge, on the detector efficiency and on the collected
gamma-ray yield. Hence, the uncertainty budget presented in Table 4.29 is to be expected.
TABLE 4.29. UNCERTAINTY BUDGET FOR THE THICKE-TARGET YIELD

MEASUREMENTS OF Al
γ-ray peak area 1%
Absolute value of the collected charge 10%
Overall uncertainty 12%
Results pertaining to the emitted yield per steradian of a pure aluminum target are shown in
Fig. 4.75. In the case of the CTN/IST laboratory [4.78], yields were also measured for (a)
Al2TiO5 and (b) Al2 O3, both of which were then converted to 100% Al, taking into
consideration the stoichiometry and the stopping power. These converted yields are in
agreement with the ones obtained for (c) a pure Al target.
164
FIG. 4.75. Thick target yields corresponding to the 844 keV γ-line of the 27Al(p,pγ1–0)27Al reaction,
emitted by a pure Al target, as a function of proton energy. Data obtained from the CRP:
a) measurements performed with Al2TiO5 and then converted to 100% Al [4.78], b) measurements
performed with Al2O3 and then converted to 100% Al [4.78], c) measurements with a pure Al target
[4.78], d) [4.101], e) [4.103] and f) calculated thick target yield obtained from a combination of
measured cross-sections (see text).
In the common energy range, Jesus et al. [4.78] values are in agreement with Chiari et al.
values [4.101] within 5%. Jokar et al. [4.103] data are lower by 15 to 25% than [4.78] below
2.5 MeV and around 15% lower than [4.101] near 3 MeV. The discrepancy with [4.78] may
be, in part, due to a mismatch of the energy scales, as indicated in Fig. 4.75. Chiari et al. have
also obtained yields at 0° and 45°. Considering that the uncertainty of the yields, at a given
angle relative to the yields at other angles, depends only on the relative uncertainty of the
detection efficiency of the different detectors, these measurements can be considered to be in
agreement, as expected for an isotropic line.
A combined excitation function was produced by adopting the cross-sections of Jesus et al.
[4.78] up to 3.0 MeV and those of Chiari et al. [4.101] from 3 to 4.1 MeV (Figs. 4.67–4.69).
This combined excitation function was introduced in the ERYA code (for details see Section
5) in order to obtain thick-target yields for a pure aluminum target, from 1.5 to 4.1 MeV
proton energy. These results are also shown as a solid black line in Fig. 4.75. As can be seen
from the figure, the calculated yields agree with the experimental results within the
experimental uncertainties.

Several previous measurements of thick-target yields were found. The ones of Anttila et al.
[4.1] and Kiss et al. [4.2] were included in the Handbook of Ion Beam Analysis [4.106].
These results, together with the ones by Deconninck et al. [4.107], are now in the IBANDL
database. Figure 4.76 displays these previous results with the ones measured in this CRP.
Deconninck’s results seem to lay in-between the results from [4.78] and the results below 2.5
MeV from [4.103]. They are 15% higher than [4.103] around 3 MeV and are in agreement
with [4.101]. Kiss et al. [4.2] results at 2.4 MeV are 18% higher than the results from [4.78]
165
(or there is an energy scale mismatch) and at higher energies his results are 10% higher than
the ones by [4.101]. The result from Anttila [4.1] at 1.7 MeV is 10% higher than the [4.78]
value.
FIG. 4.76. Thick target yields corresponding to the 844 keV γ-line, emitted by a pure Al target as a
function of proton energy. Data obtained in this CRP (from a to e, the same as for Fig. 4.75) and from
the literature: f) [4.107], g) [4.1] and h) [4.2].

Taking into consideration the uncertainty of the yield results mentioned above, all new and
old thick-target yield results are within reasonable agreement, apart from possible
discrepancies in the energy scale. Hence they may be used as supplementary information in
PIGE analysis of Al, by providing (i) data for a PIGE system calibration, (ii) data for the
determination of E1/2, i.e. the energy at which the yield is half of that obtained at incident
energy, and (iii) an indication of the yield to be expected.
27
4.9.2.2. Al(p,pγ2-0)27Al: 1014 keV γ-ray

The same experiments that measured the yields of the 844 keV γ-ray line also measured the
yields of the 1014 keV γ-ray line. The comparison of the results shown in Fig. 4.77 leads to
similar conclusions as in the previous paragraphs. The data from Jokar et al. [4.103] are lower
than data from Jesus et al. [4.78] and Chiari et al. [4.101], although overall there is slightly
better agreement than in the case of the 844 keV γ-ray line. To be more precise, the data of
[4.103] are lower by 15% compared to [4.78], for data below 2.5 MeV, and around 5% lower
than [4.101] around 3 MeV. Again, the discrepancy with [4.78] values may be partly due to a
mismatch of the energy scales.
166
FIG. 4.77. Same as in Fig. 4.75, but for the 1014 keV γ-line, emitted by a pure Al target as a function
of proton energy.
Chiari et al. [4.101] have also measured the yields at 0° and 45°. These thick-target yield data
are in good agreement with the ones at 90°, indicating that any anisotropy related to the 1014
keV γ-ray line is small.
A combined excitation function consisting of data from Jesus et al. [4.78], up to an energy of
3.0 MeV, and data from [4.101], at energies from 3 to 4 MeV (Figs. 4.67–4.69), was
introduced in the ERYA analysis code (see Section 5 for details) in order to obtain thick-
target yields for a pure aluminum target, for an energy range from 1.6 to 4.1 MeV. These
results are also shown in Fig. 4.77 as a solid black line. The agreement with the experimental
results is within the experimental uncertainties.

Several previous measurements of thick-target yields for the 1014 γ-ray line were found. The
data of Kiss et al. [4.2] were included in the Handbook of Ion Beam Analysis [4.106]. These
results, together with the data of Deconninck et al. [4.107] and Savidou et al. [4.17], are
available in the IBANDL database. Figure 4.78 shows a comparison of these previous thick-
target yield data with the data measured in this CRP.
Similar conclusions to those regarding the 844 keV γ-ray line may be drawn from the results
of Kiss et al. [4.2]. The data of Deconninck et al. [4.107] are now lower than the data of Jokar
et al. [4.103] by around 15%, for energies lower than 2050 keV; however they converge at
higher energies. The data of Savidou et al. [4.17] are in good agreement with the data of Jesus
et al. [4.78] at low energy, however they become increasingly higher than the data of Chiari et
al. [4.101] with increasing proton energy (10 % at 3 MeV and 25% at 4 MeV).
167
FIG. 4.78. Thick target yields corresponding to the 1014 keV gamma-line, emitted by a pure Al target
as a function of proton energy. Data obtained in this CRP are from a to e, the same as for Fig. 4.77,
and previous measurements from: f) [4.107] (90°), g) [4.2] (55°) and h) [4.17] (90°).

The same conclusions drawn for the 844 keV γ-ray line apply here.
27
4.9.2.3. Al(p,αγ1–0)24Mg: 1369 keV γ-ray

Along with determining yields related to the 844 and 1014 keV γ-lines, two of the laboratories
also determined yields of the 1369 keV γ-line (Fig. 4.79). Within the energy range 2500–2800
keV, results from Jokar et al. [4.103] are around 10% lower than the results from Chiari et al.
[4.101], however the discrepancy rises to 26% at 2950 keV.
Cross-section data were combined to produce an excitation function for the 27Al(p,αγ)24Mg
reaction which was then introduced in the ERYA code (see Section 5) in order to obtain thick-
target yields for a pure aluminum target, from 1.7 to 3.05 MeV proton energy. The combined
data used the measurements of [4.102] for energies up to 2.940 MeV and of Zamboni et al.
[4.79] from 2.940 MeV up to 3.05 MeV (Fig. 4.73 and as a solid black line in Fig. 4.79). The
ERYA values are larger than the measured thick-target yields from [4.103] by 50–80% in the
region from 1.8 to 2.4 MeV. From 2.4 MeV, this difference decreases with increasing energy.
From 2.6 MeV, ERYA calculated yields agree with the ones measured by [4.101] within 10%.
168
1.0E+07
1.0E+06
Yield (Ng/mC/sr)
1.0E+05
a
1.0E+04 b
c
1.0E+03
1.0E+02
1450 1950 2450 2950 3450 3950
Proton Energy (keV)
FIG. 4.79. Thick target yields per steradian corresponding to the 1369 keV γ-line, emitted by a pure Al
target as a function of proton energy. Data obtained in this CRP: a) [4.101] (45°), b) [4.103] (90°),
and c) calculated thick target yield using the ERYA code (see text and Section 5 for more details).

Several previous measurements were found in the literature. The thick-target yields of Anttila
et al. [4.1] and Kiss et al. [4.2] have been included in the Handbook of Ion Beam Analysis
[4.106]. These results together with results of Deconninck et al. [4.107] are also available in
the IBANDL database. In Fig. 4.80, the previous results are compared with the new measured
thick-target yields.
The results in Fig. 4.80 suggest that, at energies lower than 2000 keV, there is a difference in
energy scale between Deconninck et al. [4.107] and Jokar et al. [4.103] of about +20 keV,
which decreases with energy. The results of Anttila [4.1] and Kiss et al. [4.2] at 2400 keV
agree with the data of [4.103], if a +10 keV is applied to Jokar’s energy scale (or -10 keV to
the one of Anttila and Kiss). Kiss’s result [4.2] at 3100 keV is 20% higher than Chiari’s
[4.101], but there is fair agreement at 3800 keV.
169
FIG. 4.80. Thick target yields corresponding to the 1369 keV γ-line emitted by a pure Al target as a
function of proton energy. Data obtained in this CRP a) [4.101] (45°), b) [4.103] (90°) and found in
the literature c) [4.107] (90°), d) [4.1] (55°) and e) [4.2] (55°).
(c) Conclusions
The discrepancy between ERYA (see Section 5) calculated yields and the results from Jokar
et al. [4.103], at energies between 1.8 and 2.4 MeV, point to an overestimation of the
measured cross-sections measured by [4.102]. Although it is not recommended, if a
standardless PIGE analysis is performed with the help of this γ-ray line, this overestimation
should be taken into account.
Nevertheless, in spite of the above remarks, due to the fair agreement among new and old
yield results, these measurements are still useful for PIGE analysis of Al as they provide (i)
values for a PIGE system calibration, (ii) a determination of E1/2, i.e. the energy at which the
yield is half of the one obtained at incident energy, and (iii) an indication of the yield to be
expected.
(d) Summary remarks

The following is a summary of the main findings for the individual γ-ray lines mentioned in
the previous sections:
• Good agreement between the new and previous measurements of cross-section values
for the 844 and 1014 keV γ-ray lines of the 27 Al(p,p′γ)27Al nuclear reaction;
• Fair agreement between calculated values of combined excitation functions, which are
obtained from [4.78] for values between 1490 to 3000 keV and from [4.101] for values
between 3000 to 4100 keV, with measured thick-target yields relative to the same γ-
lines;
• Possiblity to perform PIGE analysis by inserting these composed excitation functions
into a PIGE analysis code such as ERYA (see Section 5) or other similar codes.
170
From these findings, it can be concluded that there is sufficient consistent experimental
information available for these γ-ray lines in order to perform a complete and comprehensive
theoretical evaluation.
4.10. SILICON
4.10.1. Gamma-ray production cross-sections of silicon from proton-induced nuclear

reactions
Silicon can be analysed by the PIGE technique by employing either the 28Si(p,p′γ)28Si
reaction or the 29Si(p,p′γ)29Si reaction, which produce the 1779 keV and 1273 keV γ-rays,
respectively. Within the frame of this CRP, new measurements were performed to determine
cross-section data for silicon analysis via PIGE, as well as the corresponding thick-target γ-
ray yields. All the new data, as well as the older data found in the literature, have been
uploaded in IBANDL.
4.10.1.1. 28Si(p,pγ1–0)28Si: 1779 keV γ-ray and 29Si(p,pγ1–0)29Si: 1273 keV γ-ray
New measurements in the energy range of 3.6 – 6.9 MeV

Cross-sections for the proton-induced nuclear reactions of 28Si(p,p′γ)28Si and 29Si(p,p′γ)29Si
have been determined in the proton energy range 3.6–6.9 MeV, at a detection angle of 55 o
[4.63]. The proton beam was generated by the 5 MV belt-driven tandem accelerator of the
University of Helsinki. The 14N(p,p′γ)14N reaction with resonances at 3903, 3996, and
5937 keV were exploited in the calibration of the absolute proton beam energy at 4–6 MeV,
with precision better than 1 keV. The calibration procedure is described in detail in [4.65].
The γ-rays were detected with a 38% efficient HPGe germanium detector positioned 20 mm
from the target. The detection angle was fixed to 55° relative to the beam direction. The
absolute efficiency curve for the detector was determined by using Eu-152, Co-60 and Co-56
calibration sources placed at the position of the target. For more experimental details, such as
charge collection and target preparation, see Section 4.5.1. The same thin self-supporting
Si3N4 membrane targets were used as in the corresponding nitrogen cross-section
measurements.
The selected membrane thickness of 100 nm was chosen to ensure sufficient counting rate of
the γ-ray detector and still enable the membrane to be considered a thin target for the present
experiments. The estimated energy loss of 4 MeV protons in the Si3N4 target is ~2.5 keV and
the energy straggling is ~3 keV, based on SRIM calculations. The sample composition and
areal densities were determined by ERDA.
The reaction excitation functions were measured with energy steps of 100 keV, except for the
energy regions provided in Table 4.30, where the energy steps were narrower. The collected
charge for each measurement point was ~12 µC. Carbon buildup at the membrane surface was
regularly checked for and reduced by frequently changing the irradiation spot on the
membrane.
171
TABLE 4.30. MEASURED ENERGY POINTS FOR THE MEASURED CROSS-SECTIONS OF
1779 keV 28Si(p,p′γ)28Si AND 1273 keV 29Si(p,p′γ)29Si
Energy interval (MeV) Energy step (keV)
3.871 - 4.016 5
4.016 - 4.086 10
4.850 - 4.960 10
4.960 - 5.380 30
5.540 - 5.600 10
5.912 - 5.957 5
6.110 - 6.192 20
6.540 - 6.840 20
The obtained cross-section data are presented in Figs. 4.81 and 4.82. The possible sources of
error in the cross-section values are:
• Absolute efficiency of the γ-ray detector. This is mainly due to the uncertainties of the
calibrated radioactive source activities;
• Determination of the absolute value of the collected charge. Despite appropriate
electron suppression, systematic deviations may occur;
• The determination of the target atomic thickness by ERDA;
• Accuracy of available stopping power data as provided by SRIM.
The uncertainty budget is summarized in Table 4.31.
TABLE 4.31. UNCERTAINTY BUDGET OF CROSS-SECTION MEASUREMENTS OF 1779 keV

28
Si(p,p′γ)28Si AND 1273 keV 29Si(p,p′γ)29Si
Beam energy spread ~2.5 keV
Energy straggling in target ~3 keV
Charge collection 3%
Peak unfolding including dead time correction 1.5% at resonance; 3 % off-resonance
Detection efficiency x solid angle 5%
Target stoichiometry 0.5%
Cross-section uncertainty: 10% at resonance; 13% off-resonance
The proton-induced thick-target γ-ray yields were measured at an angle of 55o relative to the
beam direction at 4.0, 4.5, 5.0, 5.5, 6.0 and 6.5 MeV using a thick Si3N4 target. In the thick-
target yield measurements, the collected charge was measured directly from the sample using
electron suppression. The uncertainty budget of the thick-target γ-ray yield measurements is
provided in Table 4.32.
172
TABLE 4.32. UNCERTAINTY BUDGET OF THICK-TARGET YIELD MEASUREMENTS OF
1779 keV 28Si(p,p′γ)28Si AND 1273 keV 29Si(p,p′γ)29Si
Peak unfolding including dead time correction 1%
Detection efficiency x solid angle 5%
Target stoichiometry 0.5%
Thick-target yield uncertainty 10%
(a) New measurements in the energy range of 2.0 – 3.2 MeV

In the framework of this CRP, differential cross-sections for γ-ray emission from the
28
Si(p,p′γ)28Si (Eγ = 1779 keV) and the 29Si(p,p′γ)29Si (Eγ = 1273 keV) nuclear reactions were
measured in the energy range of 2.0–3.2 MeV and 2.0–3.0 MeV, respectively, by Jokar et al.
[4.108]. The experiments were performed using the 3 MV Van de Graaff accelerator of the
Nuclear Science and Technology Research Institute (NSTRI) in Tehran. The excitation
functions were measured with energy steps of 1–2 keV in resonance regions and variable
energy steps of 10–40 keV in off-resonance regions. The proton beam energy was calibrated
using the 991.9 keV resonance of the 27 Al (p,γ)28Si reaction and the 1880.4 keV threshold
energy of the 7Li(p,n)7Be reaction. Gamma-rays were detected by a 50% P-type HPGe
detector located at 90° with respect to the incident beam direction at a distance of 51.9 mm
from the target center and subtending an angle of 65°. The absolute efficiency curve of the γ-
ray detector was determined using 133Ba, 152Eu, 137Cs, 60Co and 241Am calibration radioactive
sources, which were placed at the exact position of the target. Backscattered protons were
detected simultaneous with γ-rays using an ion implanted Si detector placed at a scattering
angle of 165°, with a solid angle of 0.746 msr. The targets were created from two films of
natural SiO evaporated onto thin self-supporting Ag layers. The stoichiometry and the
thickness of the target films were determined by RBS measurements with a 2 MeV alpha-
beam. By simulation of the RBS spectra with the SIMNRA code, composition of the SiO
layer was found to be 45% natural Si and 55% O. The thickness of the SiO film of the first
target was found to be 25.6 ± 1.8 µg/cm2. The first target was used in the 2.0–3.0 MeV
incident beam energy range, while the second one, consisting of 53 ± 4 µg/cm2 SiO, was
employed in the beam energy range of 3.0–3.2 MeV. Based on SRIM calculations, the proton
beam energy loss in SiO layers was found to be 4.3–3.2 keV and 6.5–6.2 keV for the first and
second targets, respectively. The target was oriented inside the reaction chamber so that the
incident beam direction made an angle of 45° with respect to the normal of the target. The
accumulated beam charge was chosen to be between 5 and 400 µC at each energy point,
depending on the statistics. In this way, statistical errors in the on-resonance and off-
resonance regions were found to be 1–3% and 5–10%, respectively. In order to remove large
systematic uncertainty due to the direct measurement of absolute value of the collected beam
charge, RBS of the Ag film was used for normalization of the cross-sections. To avoid
excessive carbon build up on target, the beam spot position was changed by fine adjustment
of target location several times during the measurements. The systematic uncertainty budget
is summarized in Table 4.33.
173
TABLE 4.33. SYSTEMATIC UNCERTAINTY BUDGET: CROSS-SECTION MEASUREMENTS
OF THE 1779 keV 28Si(p,p′γ)28Si AND 1273 keV 29Si(p,p′γ)29Si
The γ-ray detector absolute efficiency 7%
The stoichiometric ratio 4%
Rutherford cross-section on Ag 1%
Cross-section uncertainty 9%
The thick target γ-ray yields for the 28Si(p,p′γ)28Si (Eγ = 1779 keV) and the 29Si(p,p′γ)29Si (Eγ
= 1273 keV) nuclear reactions were also measured at the laboratory angle of 90°, from 2.38–
3.0 MeV in 10 keV energy steps and 2.2–3.0 MeV with 10 keV energy steps, respectively
[4.103]. The target was a thick Si plate, in a single-element form. Additionally, a 70 × 1015
atoms/cm2 Au layer was evaporated on the surface of the target in order to measure the
collected charge. The experiments were performed in the same experimental set-up used for
cross-section measurements. The uncertainty budget of the thick-target γ-ray yields is
presented in Table 4.34.
TABLE 4.34. UNCERTAINTY BUDGET: THICK-TARGET YIELD MEASUREMENTS OF 1779

keV 28Si(p,p′γ)28Si AND 1273 keV 29Si(p,p′γ)29Si
The gamma-ray detector absolute efficiency 7%
Stopping power 4%
Thick-target yield uncertainty 9%

The available PIGE cross-section data in the literature for silicon is scarce. Only the studies of
Dyer et al. [4.52] and Boni et al. [4.5] can be found for the energy range covered by the
present differential cross-section measurements. The values of Jokar et al. are for proton
energies below 3.2 MeV [4.108]. The available previous cross-section values, together with
the present values for the reaction 28Si(p,p′γ)28Si, are shown in Fig. 4.81. When comparing the
different data sets, it should be kept in mind that the associated angular distributions of the γ-
rays are not isotropic [4.52]. The data of [4.5] and [4.108] are differential cross-section values
obtained at specific detection angles and in Fig. 4.81, the data have been plotted as total cross-
section values, taking into account the appropriate angular distribution determined by [4.52].
174
28Si(p,p'γ)28Si
103
Cross section [mb]
102
a
b
c
d
101
3000 4000 5000 6000 7000
Ep [keV]
FIG. 4.81. Cross-sections for production of 1.78 MeV γ-rays from the reaction 28Si(p,p′γ)28Si as a
function of proton energy. Data obtained from: a) [4.52], b) [4.63] (55°), c) [4.5] (90°), and
d) [4.108] (90°).
The work of Dyer et al. [4.52] covers the proton energy range from 3 MeV to 23 MeV. In this
study, the γ-ray angular distribution was determined by measuring yields at several detection
angles and by fitting them with a series of Legendre polynomials. The estimated uncertainty
in the Dyer et al. cross-section values is 15% for proton energies below 5 MeV and 10% for
values above 5 MeV. The experimental points are in reasonable agreement with the other
data, in the lower part of the energy range. Within the energy range 4.4–5.1 MeV, the values
of Dyer et al. are clearly lower than the new data of Marchand et al. [4.63]. In the high energy
region, the agreement becomes better.
In the study of Boni et al. [4.5], cross-section data is provided for the energy range 2.2 – 3.8
MeV, measured at a 90° angle [4.5], with energy steps of 20 keV. The data points of [4.63], in
the overlapping energy region, are in reasonable agreement with those of Boni et al. A shift of
30–35 keV in the proton energy would yield completely overlapping data points.
The present values of [4.108] range from 2 MeV to 3.2 MeV proton energy. The data has
been collected at a detection angle of 90o. A detailed comparison between the values of this
study and those of Boni et al. can be found in the original article by Jokar et al. [4.108].
The comparison shows that, in the proton energy range of 3.0–3.2 MeV, there is a very good
agreement between the measured cross-section values of [4.108] and those of Boni et al.
175
However, due to using thinner target and smaller energy steps in the resonance region, the
obtained resonance in the present data [4.108] is sharper than results from Boni et al. [4.5].
Except for the 1.78 MeV γ-line produced by 28 Si, no other cross-section data for proton
reactions with silicon, within the presently considered energy range, are available in the
literature.
The present data for the 29Si(p,p′γ)29Si reaction is shown in Fig. 4.82. The data of [4.108] are
differential cross-section values obtained at a 90o detection angle. The data has been plotted in
Fig. 4.82 as total cross-section values, assuming isotropic angular distribution.
104
29Si(p,p'γ)29Si
103
Cross section [mb]
102
1
10
a
b
0
10
2000 3000 4000 5000 6000 7000
Ep [keV]
Fig. 4.82. Cross-sections for production of the 1.27 MeV γ-rays from the reaction 29Si(p,p′γ)29Si as a
function of proton energy. Data obtained within CRP: a) [4.108] (90°) and b) [4.63] (55°).
Even though PIGE analysis of silicon is solely based on the isotopes of 28Si and 29Si, depth
profiling of silicon by nuclear reaction resonances is typically carried out by utilizing the 30Si
isotope and the Ep = 620 keV resonance of the 30Si(p,γ)31P reaction [4.109–4.111].
For silicon, the following individual thick-target γ-ray yield data points are available in the
literature: at 2.4 MeV [4.1], 3.1 and 3.8 MeV [4.2], 7 MeV [4.3], 3.05–4.09 MeV [4.17] and
2.65 – 4.13 MeV [4.112]. A more systematic approach has recently been taken by Jokar et al.
with yield values measured from 2.2 MeV to 3 MeV with energy steps of 10 keV [4.108]. In
the work of Marchand et al. [4.63], additional points have been obtained at 4.0, 4.5, 5.0, 5.5,
6.0 and 6.5 MeV. In Fig. 4.83, the present yield values for silicon are compared with available
literature data. The data of [4.2], [4.3] and [4.1] have been obtained at an angle of 55o (as is
the case of data from [4.63]), the data of [4.112] at angles of 00, 45o, 90o and those of [4.17]
176
and [4.108] at 90o. It should be noted that, in this case, the yields can be very anisotropic and
the comparison between the values is only indicative. The yields calculated from the
measured differential cross-sections shown in Figs. 4.81 and 4.82 have not been included in
the comparison, since the measured excitation functions have missed the narrow resonances
and therefore provide only rough estimates.
107
29Si(p,p'γ)29Si
Thick-target yield /µC sr
. 106
105
a
b
c
104
d
e
3
10
2500 3000 3500 4000 4500 5000 5500 6000 6500
Ep [keV]
28Si(p,p'γ)28Si
8
10
Thick-target yield /µC sr
.
107
a
6 b
10
c
d
e
105
f
g
104 h
i
103
2000 3000 4000 5000 6000 7000 8000 9000
Ep [keV]
FIG. 4.83. Comparison of the measured thick-target yields with available previous literature data
corresponding to natural isotopic abundance as a function of proton energy. Upper figure illustrates
the γ-ray line 1779 keV from the reaction 28Si(p,p′γ)28Si: a-b) [4.112] (0°, 45°, 90°), c) [4.2] (55°),
d) [4.1] (55°), and e) [4.63] (55°). Lower figure illustrates the 1273 keV line from the reaction
29
Si(p,p′γ)29Si: a) [4.108] (90°), b) [4.17] (90°), c-e) [4.112] (0°, 45°, 90°), f) [4.1] (55°), g) [4.2]
(55°), h) [4.3] (55°) and i) [4.63] (55°).

Both γ-ray lines are suitable for silicon analysis, specifically for detection of the two separate
isotopes. Since the excitation curves for both reactions indicate the presence of several narrow
resonances, more accurate data obtained with narrow energy steps and good resolution are
recommended. The available thick-target yield data is still rather limited.
177
4.10.3. Gamma-ray production cross-sections of silicon from deuteron-induced nuclear
reactions
The deuteron-induced gamma-emission is mainly applied to the analysis of light elements, but
it also simultaneously detects silicon. As previous thick-target measurements [4.18] indicate,
many states are excited, mainly in the 28Si(d,pγ)29Si reaction, which produces a rather
complicated gamma-spectrum even at low deuteron energies. The most intense γ-rays from
this reaction are the 1273 keV, 2028 keV, 2426 keV and the 4934 keV from the de-excitation
of the 1st, 2nd, 3rd and 10th excited states, respectively, to the ground state. The only available
cross-section measurements for these γ-rays were performed in the frame of this CRP [4.113].

Differential cross-sections for γ-ray emission from the 28Si(d,pγ)29Si reaction (Eγ = 1273,
2028, 2426 and 4934 keV) have been determined for deuteron energies ranging from 0.65 to
2.0 MeV, at the detection angle of 55o. The details of the experiment are the same as for
nitrogen (see Section 4.5), since the same thin self-supporting Si3N4 film target was used
[4.113].
The main source of uncertainty was the measured γ-peak area, which varied between 1% and
40% for the silicon γ-lines, depending on the incident deuteron energy and the silicon reaction
channel. Average uncertainties of the silicon γ-ray production cross-sections are 12%. The
uncertainty budget is presented in Table 4.35.
TABLE 4.35. UNCERTAINTY BUDGET: SILICON ABSOLUTE DIFFERENTIAL CROSS-

SECTION VALUES
Beam energy spread (keV) 3 keV
Charge determination 3.0 %
N number of target nuclei/cm2 2.5 %
γ-ray peak area 1–40 %
γ-ray detector absolute efficiency 2.0 %
Uncertainty 12 %
The resulting differential γ-ray cross-sections are rather small and have large uncertainties in
the measured energy region. Only the 1273 keV and 4934 keV γ-rays are suitable for PIGE
analysis.
4.10.3.1. 28Si(d,pγ)29Si reaction: 1273 γ-ray
The cross-sections for the 1273 keV γ-ray from the 28Si(d,pγ)29Si reaction are plotted in Fig.
4.84.
178
FIG. 4.84. Differential cross-sections of the 1273 keV γ-ray produced by the 28Si(d,pγ)29Si reaction at
55°, as a function of deuteron energy [4.113].
A critical assessment cannot be performed since this is the only existing measured dataset.
Yet validation can be performed through a comparison of the calculated yields from the
differential cross-section values with the only existing measured thick-target gamma-yield. In
Fig. 4.85, a comparison with the thick-target yields of Elekes et al. [4.18] is presented. The
comparison shows that there is a large discrepancy between the calculated and measured yield
data, below 1300 keV. Above 1300 keV, the calculated yields gradually converge with the
measured thick-target gamma-yield data. The discrepancy remains within the 15 %
uncertainty band of the measured curve in the entire energy range.
179
FIG. 4.85. Comparison of the thick target yields of the 1273 keV γ-ray produced by the 28Si(d,pγ)29Si
reaction as a function of deuteron energy. Data obtained from a) [4.18] (135°) and b) [4.113] (55°).
4.10.3.2. 28Si(d,pγ)29Si reaction: 4934 γ-ray
The 4934 keV excited state has a 3/2 spin value, which results in the angular distribution
being anisotropic (although the anisotropy is not very large [4.113]). The 1.2 fs lifetime of the
excited state results in a Doppler broadening of the shape of the γ-ray. The measured γ-ray
production cross-sections are plotted in Fig. 4.86 as a function of deuteron energy. The cross-
section values increase rapidly, but are still quite low in this deuteron energy region. Because
of the high energy of this γ-ray, it can be identified easily in a bulk spectrum containing
several elements.
180
FIG. 4.86. Differential γ-ray production cross-sections of the 4934 keV γ-ray produced by the
28
Si(d,pγ)29Si reaction at 55° as a function of deuteron energy [4.113].
As in the case of the 1273 keV γ-ray, a comparison of the calculated yields from the
differential cross-section values with the only existing measured thick-target γ-yield can help
validate these cross-section values. In Fig. 4.87, a comparison with the thick-target yields of
Elekes et al. [4.18] is presented. The comparison shows a good agreement between the
calculated and measured thick-target gamma-yields. The discrepancy between the calculated
and measured yield data remains within the 15 % band uncertainty for the measured curve,
along the whole deuteron energy region.
FIG. 4.87. Comparison of the thick target yields of the 4934 keV γ-ray produced by the 28Si(d,pγ)29Si
reaction as a function of deuteron energy. Data obtained from: a) [4.18] (135°) and b) [4.113] (55°).
181
Although both γ-ray lines have intensities large enough for silicon analysis using deuteron
energies at or above 1.8 MeV, the application of the 4934 keV γ-ray is more highly
recommended because of its easy identification in a bulk spectrum containing several
elements. Because of the absence of any other experimental data, additional measurements
with both thin and thick targets are also recommended.
4.11. PHOSPHORUS
4.11.1. Gamma-ray production cross-sections of phosphorus from proton-induced

nuclear reactions
Proton-induced nuclear reactions on the single stable isotope of Phosphorus, 31P, produce
several γ-rays. Among them, the 1266 keV γ-ray, which is emitted following the inelastic
scattering of protons by phosphorus, is the most useful one for analysis of in-depth
homogeneous samples, since it has the highest yields. It is also expected to have an
anisotropic angular distribution due to the spin 3/2 of the first excited state of 31P. The other γ-
ray lines produced by the same process have higher energies and the detectors respond with
lower efficiencies. Other γ-rays are also produced by proton radiative capture reactions.
However, since the cross-sections of these capture reactions are much smaller than the
inelastic scattering cross-sections and have lower sensitivity, the inelastic scattering cross-
sections are preferred for in-depth homogeneous sample analysis. In contrast, for depth
analysis, proton capture is preferred since it exhibits narrow and rather isolated resonances,
which are more suitable for depth profiling purposes.
Regarding the available data for γ-ray production, cross-sections from phosphorus proton
inelastic scattering and capture reactions are scarce in the energy region of interest. In view of
this situation, it was decided to perform measurements of γ-ray production cross-sections of
phosphorus from proton inelastic scattering.
4.11.1.1. 31P(p,pγ1–0)31P: 1266 keV γ-ray

Measurements of 31P(p,pγ1–0)31P reaction were performed at CTN/IST (Lisbon) and at the
NSTRI (Tehran). The experimental conditions for the different experiments are presented
below.
CTN /IST (Lisbon)
The proton beams were accelerated by a 3 MV Tandetron accelerator, with energies in the
range of 2.5 to 4.0 MeV. Proton beam currents to the order of 100 nA were employed. The
proton energy was calibrated by the 1645.1 keV and 1930.7 keV resonances of the
23
Na(p,p′γ)23Na reaction and by the 3470 keV resonance of the 16O(p,p)16O reaction. The
reaction chamber was electrically insulated from the beam line and worked together with the
target holder and beam stopper as one Faraday cup for beam charge collection. Scattered
protons at 155° with respect to the incident beam were detected by Camberra PIPs detectors
with active area of 50 mm2 and a 100 mm depletion layer, located in Cornell geometry at 88
mm from the target, with solid angles defined by apertures of 6.00 ± 0.05 mm. Gamma-rays
were detected by a 45% Ge(HP) detector located at 130° with respect to the incident beam
direction and at a distance of 55.5 mm from the target. The absolute efficiency versus γ-ray
182
energy was obtained by measuring the yields of γ-rays produced by ion sources (133Ba and
152
Eu) calibrated in activity as well as by Monte Carlo simulations using GEANT and
PENELOPE.
The target was made from a thin film of Ca3(PO4)2 evaporated onto a thin self-supporting Ag
film. RBS of the Ag film was used to normalize the cross-sections. The uncertainty budget is
presented in Table 4.36.
TABLE 4.36. UNCERTAINTY BUDGET FOR THE CROSS-SECTION MEASUREMENTS OF

1266 keV 31P(p,pγ1–0)31P – CTN/IST
P/Ag atomic ratio 10%
Proton beam energy 0.1%
NSTRI (Tehran)
The cross-sections were measured [4.12] in the proton energy range of 1886–3007 keV at the
laboratory angle of 90º, employing a thin Zn3P2 target evaporated onto a self-supporting C.
Beam intensities of the order of 50 nA were used. The accelerator beam energy was calibrated
using the 991.88 keV resonance of the 27Al(p,g)28Si reaction and the 1880.44 keV threshold
energy of the 7Li(p,n)7Be reaction. The γ-rays and backscattered protons were detected
simultaneously. A 50 % HPGe detector placed at 51.9 mm collected the γ-rays while an ion
implanted Si detector placed at a scattering angle of 165º detected the scattered particles. The
uncertainty budget may be found in Table 4.37.

1266 keV 31P(p,pγ1–0)31P – NSTRI
γ-ray peak area 1-3%
P/Ag atomic ratio 4%
Particle detector solid angle 2%
Figure 4.88 presents values obtained by the referred measurements. These values may be
found in the IBANDL database: CTN/IST – Jesus et al. [4.12] from 2621 to 3950 keV in
steps from 1 keV (on resonance) to 10 keV (off resonance); NSTRI – Jokar et al. [4.114] from
1886 to 3007 keV in steps from 2 (on resonance) to 10 keV (off resonance).
183
In the common energy range, the energy scales of the two new measurements are in
agreement within 1 keV. The cross-sections differ by an average of 12%, well within the
quoted uncertainties.

A review of the existing literature only produced the measurements of Boni et al. [4.5] (data
available in IBANDL). The cross-section data were measured in the proton energy range from
2648 to 3799 keV, in steps ranging from 15 to 20 keV. Due to the large energy step, some of
the resonance maxima may have been missed, which hinders comparison with results
obtained in this CRP. Nevertheless, it can be concluded that the results from Boni are higher
by 10 to 20% (see Fig. 4.88), considering the quoted uncertainties.
1.6E+01
a
b
Cross Section (mb/sr)
1.2E+01
c
8.0E+00
4.0E+00
0.0E+00
2600 2800 3000 3200 3400 3600 3800 4000
Proton Energy (keV)
FIG. 4.88. Comparison of cross-sections of the 31P(p,pγ1–0)31P reaction obtained within the CRP:
a) Jesus et al. [4.12] (130°), b) Jokar et al. [4.114] (90°) with previous results c) Boni et. al. [4.5]
(90°) in the proton energy range 2640 – 3950 keV.

From the above comparison of new and old measurements, it can be concluded that there is an
overall agreement within the experimental uncertainty. However, to be fully confident in the
results, more measurements should be made to confirm the structure that was observed, to
identify possible missed structures and to consolidate the absolute values of the excitation
function.
184
4.11.2.1. 31P(p,pγ1–0)31P: 1266 keV γ-ray

Thick-target yields were also measured during this CRP by Jesus et al. [4.12] and Jokar et al.
[4.114] at the same laboratories and with the same experimental conditions described in the
previous section. Thick-target yield results depend on the absolute value of the collected
charge, on the detector efficiency and on the collected γ-ray yield. The uncertainty budget for
the yields of [4.12] is presented in Table 4.38. In this case, the yields were measured for
Ca3(PO4)2 and then converted to 100% phosphorus, taking into consideration the
stoichiometry and the stopping power.
TABLE 4.38. UNCERTAINTY BUDGET FOR THE THICK-TARGET YIELD MEASUREMENTS

OF 1266 keV 31P(p,p′γ1–0)31P
γ-ray peak area 1%
Absolute value of the collected charge 10%
The measured thick-target yields per steradian of a pure phosphorus target are shown in Fig.
4.89.

Several previous measurements of thick-target yields for phosphorus were found in the
literature. The data of Anttila et al. [4.1] and Kiss et al. [4.2] were included in the Handbook
of Ion Beam Analysis [4.106]. These results, together with the data of Demortier et al. [4.115]
and Savidou et al. [4.17], are now available in the IBANDL database. Figure 4.89 shows a
comparison of the previous thick-target yield data with the new data measured in this CRP.
The results are in good agreement, within uncertainties, with the new data of Jesus et al.
[4.12] and Jokar et al. [4.114]. The data of Jesus et al. becomes progressively larger than the
results of Savidou et al. until finally reaching a maximum disagreement of around 22% at the
highest energies. However, this 22% disagreement is still within the quoted uncertainties.
The excitation function of Jesus et al. [4.12], shown in Fig. 4.88, was introduced in the ERYA
analysis code (see Section 5) in order to obtain thick-target yields for a pure phosphorus
target, from 2.7 to 4.0 MeV proton energy. These results are also shown in Fig. 4.89 as a solid
black line. The agreement with the experimental thick-target yields is within the experimental
uncertainties.
185
1.0E+08
a
1.0E+06
Yield (µC.sr)-1 b
c
1.0E+04
d
e
1.0E+02 f
g
1.0E+00
1740 2240 2740 3240 3740 4240
Proton Energy (keV)
FIG. 4.89. Thick target yields, per steradian, corresponding to the 1266 keV γ-line, emitted by a pure
phosphorus target as a function of proton energy. Data obtained in previous measurements:
a) Demortier et al. [4.115] (90°), b) Savidou et al. [4.17] (90°) c) Kiss et al. [4.2] (55°), d) Anttila et
al. [4.1] (55°), f) Jokar et al. [4.114] and in this CRP: e) Jesus et al. [4.12] (130°). The new
measurements were done with Ca3(PO4)2 and then converted to 100% phosphorus. The solid line g
represents calculated values by the ERYA code (see text).
(c) Conclusions
Considering the uncertainty on the yield measurements, all the new and old results seem to be
in fair agreement, apart from possible discrepancies in the energy scale. Hence they may be
used in PIGE analysis of phosphorus as complementary data for (i) a PIGE system
calibration, (ii) the determination of E1/2, i.e. the energy at which the yield is half of the one
obtained at incident energy, and (iii) an indication of the yield to be expected.

Although it has been shown that there is (i) good agreement between the two sets of cross-
section data for the 31P(p,p′γ)31P nuclear reaction producing the 1266 keV γ-ray, (ii) good
agreement between calculated thick target yields using the cross-section data of Jesus et al.
from 2600 to 4000 keV with measured thick-target yields for the same gamma line and (iii)
the possibility to perform PIGE analysis by inserting this excitation function into the ERYA
code (see Section 5) or similar code, it can be argued that a theoretical evaluation is also
necessary, followed by further validation by benchmark measurements.
4.12. SULPHUR
4.12.1. Gamma-ray production cross-sections of sulphur from proton-induced nuclear

reactions
Most of the IBA techniques are well suited for the proper quantification of Sulphur. The one
most commonly used, because of its multi-elemental capabilities, is PIXE. However, only a
few of the nuclear techniques can provide depth profile information and, in most cases, only
186
when sulphur is present in suitable matrices. On the other hand, resonant PIGE can fill this
gap by providing accurate depth profile quantification, due to the strong resonances that are
present in the 32S(p,p′γ)32S reaction. However, since recently published data disagree on the
energy and strength of these resonances, a new set of measurements was considered
necessary.
4.12.1.1. 32S(p,pγ1–0)32S : 2230 keV γ-ray

The 32S(p,p′γ)32S reaction was studied at the 5.5 MV Tandem Accelerator installed at the
Institute of Nuclear and Particle Physics, NCSR “Demokritos” by Lagoyannis et al. [4.116].
The energy calibration of the accelerator was performed at Ep = 991.9 and 1746.9 keV using
the resonances of the 27Al(p,γ)28Si and 13C(p,γ)14N reactions, respectively.
Two targets were used: a thin MoS2 film deposited on a tantalum backing and a thick MoS2
pellet. The primary goal of using the thin target was to measure the off-resonance cross-
section of the reaction, while the thick target was used to determine the exact energy of the
resonances and their strength. The targets were placed at the centre of a cylindrical chamber,
which also acted as a Faraday cup. The beam charge was collected using a current integrator,
while a suppression voltage of +300 volts was applied onto a tantalum collimator placed ~1
meter before the chamber.
The produced γ-rays were detected using four HPGe detectors placed at 0ο, 55ο, 90ο and 165 ο
with respect to the beam direction and at a distance of ~ 25 cm from the target covering an
angular acceptance of ± 10o. The detectors were then rotated by 15o to a set of four new
angles, i.e. 15o, 40o, 105o and 150o, in order to measure the angular distribution of the 2230
keV γ-ray.
The energy and efficiency calibration of the HPGe detectors were performed using a
calibrated point source of 226Ra.
The statistical error from the integration of the accumulated spectrum peaks did not exceed
5% and the overall systematic error was estimated at 8%, taking into account the uncertainties
in the beam charge collection, efficiency calibration and target thickness. An analytical
breakdown of the uncertainties is shown in Table 4.39.
TABLE 4.39. UNCERTAINTY BUDGET FOR PROTON-INDUCED SULFUR REACTIONS
Beam charge 3.5%
The measured relative yield at 55o, with respect to the proton beam, is presented in Fig. 4.90.
187
FIG. 4.90. Relative yield of the 2230 keV γ-ray at 55oas a function of proton energy [4.116].
The angular distributions of the 2230 γ-ray on the first four resonances, along with their fit
with Legendre polynomials, are shown in Fig. 4.91. Because of the complexity of the
excitation function of the 32S(p,p′γ)32S reaction, which exhibits numerous overlapping
resonances, no cross-sections were derived.
FIG. 4.91. Angular distributions of the first four resonances at (a) 3094 keV (b) 3195 keV (c) 3379
keV and (d) 3714 keV of the 2230 keV γ-ray [4.116].

The only existing data in the literature, except from those referring to the energy and the
strength of the first three resonances, are thick-target yields by Kiss et al. [4.2] shown in Fig.
4.92.
188
32
FIG. 4.92. Thick target yields of the 2130 keV γ-ray produced by the S(p,p′γ)32S reaction as a
function of proton energy at 55° [4.2].
In Table 4.40, a compilation of the existing data regarding the resonances is presented.
TABLE 4.40. RESONANCE PARAMETERS OF THE 32S(p,p′γ)32S REACTION

Resonance energy Cross-section at 90o Cross-section at 0o
(keV) (mb.keV/sr) (mb.keV/s
Reference
1st 2nd 3rd res. 1st 2nd 3rd 1st 2nd 3rd
res. res. res. res. res. res. res. res.
[4.117] 3089 3379 3717 8 30 35 18 ± 1 64 ± 3 70 ± 6
[4.118] 3095 3379 3716 10.88 41.84 48.10 — — —
[4.119] 3094 3379 3716
[4.116] 3094 3379 3714 6.2 ± 19.0 ± 20.5 ± 8.0 ± 27.5 ± 30.7 ±
0.4 1.1 1.2 0.5 1.6 1.8
From Table 4.40, it is clear that regarding the exact energies of the three resonances, only the
position of the second one is common between the various authors. Moreover, the measured
strength of the resonances varies in the literature: the strengths from Lagoyannis et al. [4.116]
have the lowest values while those from Tsartsarakos et al. [4.118] have the highest values.
The differences could be attributed to the angular acceptance of each experiment and to the
strong angular dependence of the resonances. The experiment of Tsartsarakos et al. [4.118]
was performed at an angle of 90o ± 30o, Rao et al. [4.116] at 0o and 90o ± 20o, and Lagoyannis
et al. [4.116] at 0o and 90o ± 10o.

Since only one thick-target yield dataset exists in the literature, it is not possible to make a
recommendation for these cross-sections. In the case of the resonances that are important for
depth profiling, only the energy of the second resonance can be recommended since all the
available measurements agree. Finally, for the strength of the resonances, the values
189
suggested by Lagoyannis et al. [4.116] are recommended, since they were performed with a
narrow angular divergence and thus are less sensitive to the angular distribution.
4.12.1.2. 32S(p,γ1–0)33Cl, 32S(p,γ6–1)33Cl, 34S(p,γ1–0)35Cl and 34S(p,pγ1–0)34S: 811, 2035, 1219

and 2127 keV γ-rays, respectively

The 811 keV, 2035 keV and 1219 keV and 2127 keV γ-rays are emitted by the 32S(p,γ1–0)33Cl,
32
S(p,γ6–1)33Cl, 34S(p,γ1–0)35Cl and 34S(p,pγ1–0)34S reactions, respectively. For these γ-rays,
only thick-target yields exist in the literature, as measured by Kiss [4.2] and Anttila et al.
[4.1]. Moreover, the data of Kiss et al. [4.2] have been normalized to those of Anttila et al. at
Ep = 2.4 MeV, which means that these measurements are not independent. A compilation of
the aforementioned data can be found in Fig. 4.93.
FIG. 4.93. Thick target yields measured by proton induced reactions from Sulphur at an angle of 55°
as a function of proton energy. Data obtained from: (a) 811 keV [4.1], (b) 811 keV [4.2], (c) 2035
[4.1], (d) 1219 keV [4.1], (e) 1219 keV [4.2] and (f) 2127 [4.2].

Since the two thick-target yield measurements are not independent, it is recommended that
additional thick-target yield measurements be made.
190
4.12.2. Gamma-ray production cross-sections of sulphur from deuteron-induced nuclear
reactions
4.12.2.1. 32S(d,pγ1–0)33S: 841 γ-ray
As in the case of proton capture reactions on sulphur, there is only one measurement of thick-
target yields in the literature, which was made by Elekes et al. [4.18]. The yield of the 841
keV γ-ray emitted by the 32S(d,pγ1–0)33S reaction is shown in Fig. 4.94.
FIG. 4.94. Thick target yield of the 841 keV γ-ray produced from the 32S(d,pγ1–0)33S reaction at 135° as
a function of proton energy [4.18].
(a) Recommended data

Since there is only one dataset available, no recommendation can be made. These
measurements are of no interest for the analysis of sulfur, although thay might be of interest
for multi-elemental bulk analysis of material containing sulfur. Nevertheless, they are
considered low priority.
4.13. TITANIUM
4.13.1. Gamma-ray production cross-sections of titanium from proton-induced nuclear

reactions
4.13.1.1. 48Ti(p,γ1–0)49V + 49Ti(p,nγ1–0)49V: 90.6 keV γ-ray
The 90.6 keV γ-ray from 48Ti(p,γ)49V and 49Ti(p,nγ)49V reactions have large yields and are
therefore suitable for PIGE analysis [4.120]. The aim of the Ti work performed for the CRP
191
was to measure differential cross-sections for the production of 90.6 keV γ-rays from the
48
Ti(p,γ)49V and 48Ti(p,γ)49V + 49Ti(p,nγ)49V reactions for proton energies ranging between
1.0 and 3 MeV at the laboratory angle of 900 using thin, isotopic, natural targets.

Measurements were carried out with the Van De Graff accelerator of Kharkov Institute of
Physics and Technology (KIPT). The main parameters of the accelerator are the following:
acceleration voltage 0.5–3.5 МeV, energy spread 0.07 %, ion beam current of the direct
output 90 µА, ion beam current after the bending magnet 20 µА, accelerated ions H+ and
He+. The accelerator energy was calibrated by means of well-known resonances in the
27
Al(p,γ)28Si reaction.
Thin TiN targets (natural Ti) on carbon backings were prepared by the ion beam assisted
deposition (IBAD) technique. The isotopic target (48Ti) on Ta backing was prepared by
electrodeposition. The thickness (at/cm2) of the targets were measured using back-scattering
spectrometry with He and H ions at energies ranging from 1.6 to 1.8 MeV.
The measurements of the low-energy γ-rays were performed with a thin HPGe-detector. The
calibration of the γ-ray detection system efficiency ε(Eγ) was carried out with the standard
133
Ba, 152Eu, and 241Am sources at the same geometry used for the cross-section
measurements.
The excitation functions for the production of 90.6 keV γ-rays from the 48Ti(p,γ)49V and
48
Ti(p,γ)49V + 49Ti(p,nγ)49V reactions for proton energies ranging between 1.0 and 3 MeV at
the laboratory angle of 900 were measured with a variable energy step (ranging from 18 to 50
keV). The measured γ-ray production differential cross-sections are presented in Figs. 4.95
and 4.96.
0.10
0.08 a
b
0.06
dσ/dΩ (mb/sr)
0.04
0.02
0.00
1000 1500 2000 2500 3000
E p(keV)
FIG. 4.95. a) The measured differential cross-section for production of the 90.6 keV γ-ray from the
reaction 48Ti(p,γ1–0)49V as a function of proton energies ranging between 950 and 2820 keV at the
laboratory angle of 900, b) σ TOT/4π from the work of [4.121]. The target enrichment was equal 0.978
for 48Ti.
192
0.3
0.2
dσ/dΩ (mb/sr)
0.1
0.0
1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200
Ep( keV)
FIG. 4.96. The measured differential cross-section for production of the 90.6 keV γ-ray produced from
the 48Ti(p,γ1–0)49V and 48Ti(p,γ1–0)49V + 49Ti(p,nγ1–0)49V reactions as a function of proton energies
ranging between 900 and 3100 keV, at the laboratory angle of 900.
The uncertainty budget is presented in Table 4.41.
TABLE 4.41. ESTIMATED UNCERTAINTY BUDGET FOR DIFFERENTIAL CROSS-SECTION

MEASUREMENTS/PARAMETERS OF Ti REACTIONS
Parameters Uncertainty
γ-ray detector efficiency 6–7%
Target thickness 5%
Integrated beam charges 1%

The total cross-section of 48Ti(p,γ)49V, for astrophysical purpose, was measured by Kennet et
al. [4.121] and is compared with the new data in Fig. 4.95. The total cross-sections of [4.121]
have been converted to differential cross-sections by dividing by a factor of 4π, assuming the
γ-ray emission is isotropic. Gosset used two strong resonances of the 48Ti(p,γ)49V reaction at
1007 keV and 1013 keV for profiling Ti implanted in M50 steel [4.122].
193
4.13.1.2. Thick target yield: 90.6 keV γ-ray
Demortier et al. [4.120] has carried out thick-target yield measurements on natural titanium,
as shown Fig. 4.97.
1000
Yield (Nγ/sr/µC)
100
10
1
400 600 800 1000 1200 1400 1600
Ep(keV)
FIG. 4.97. Gamma-ray yield of the 90.6 keV γ-ray at 90° produced from natural titanium as a function
of proton energy. Data from [4.120].
Thick-target yields were calculated from the new measured differential cross-sections. Since
cross-sections were not measured at energies below 1000 keV, it is convenient to compare the
difference of thick-target yields at the current proton beam energy and the thick-target yields
at beam energy 1000 keV (Y(Ep)–Y(1000 keV)). The differences are shown in Fig. 4.98.
194
5
10
a
Y(Ep)-Y(1000 keV), (Nγ/(µC*sr))
b
4
10
3
10
2
10
1
10
800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200
Ep(keV)
FIG. 4.98. The difference of thick target yields at current proton beam energy and thick target yields
at the beam energy 1000 keV for 90.6 γ-ray produced from natural titanium at 90° as a function of
proton energy. Data from: a) thick target yields calculated from the new measured differential cross-
sections and b) measured thick-target yields [4.120].

The excitation function for the 90.6 keV γ-ray produced by the 48Ti(p,γ)49V reaction shown in
Fig. 4.95 can be used for analytical purposes only up to the energy of 1.5 MeV, since there is
additional γ-ray yield from the 49Ti(p,nγ)49V reaction contributing at energies above 1.5 MeV.
The excitation function for the production of the 90.6 keV γ-ray from the 48Ti(p,γ)49V and
48
Ti(p,γ)49V + 49Ti(p,nγ)49V reactions, as shown in Fig. 4.96, can be used for analytical
purposes in the whole energy range. All the new data have been uploaded in IBANDL.
4.13.1.3. 46Ti(p,pγ1–0)46Ti: 889 keV γ-ray and 48Ti(p,pγ1–0)48Ti: 983 keV γ-ray
For the analysis of titanium by PIGE, the differential cross-sections for the production of the
889 keV γ-ray from the 46Ti(p,p′γ)46Ti reaction and the 983 keV γ-ray from the 48Ti(p,p′γ)48Ti
reaction can also be used. These cross-sections have been measured by Krivonosov et al.
[4.123] and are presented in Fig. 4.99. However, the magnitude of these cross-sections is
significantly smaller than that of the cross-section for production of the 90.6 keV γ-ray for
proton energy up to 2.5 MeV, as shown in Figs. 4.95 and 4.96. Therefore, the latter are
preferable for PIGE applications.
195
0.8
a
0.6 b
dσ/dΩ (mb/sr)
0.4
0.2
0.0
1.5 2.0 2.5 3.0

Ep( keV)
FIG. 4.99. Measured differential cross-sections as a function of proton energy for the a) 889 keV γ-ray
from the 46Ti(p,pγ1–0)46Ti reaction, b) 983 keV γ-ray from the 48Ti(p,pγ1–0)48Ti reaction. Data from
[4.123].
4.14. CHROMIUM
The 378 keV γ-ray produced from the 52Cr(p,γ1–0)53Mn and 53Cr(p,nγ1–0)53Mn reactions is
suitable for analytical measurements according to [4.120] and [4.122]. The aim of the CRP
was to measure differential cross-sections for the production of 378 keV γ-rays from the
52
Cr(p,γ)53Mn + 53Cr(p,nγ)53Mn reactions for proton energies ranging between 1.0 and 3 MeV
at the laboratory angle of 900 using thin natural targets. All the data have been uploaded in
IBANDL.
4.14.1. Gamma-ray production cross-sections of chromium from proton-induced nuclear

reactions
4.14.1.1. 52Cr(p,γ1–0)53Mn + 53Cr(p,nγ1–0)53Mn: 378 keV γ-ray

Measurements were carried out with the Van De Graff accelerator of KIPT (see Section 4.13).
The accelerator energy was calibrated by means of the well-known resonances produced by
the 27Al(p,γ)28Si reaction.
Thin natural Cr targets on Ta backings were prepared by the vacuum deposition technique.
The thickness (at/cm2) of the targets were measured by using back-scattering spectrometry
with He and H ions at energies ranging from 1.6 to 1.8 MeV.
The γ-ray measurement was performed using a co-axial Ge(Li)-detector. Calibration of the γ-
ray detection system efficiency ε(Eγ) was carried out with the standard 133Ba and 152Eu
sources with the same geometry used for the cross-section measurements.
196
The excitation functions for the production of the 378 keV γ-rays from the 52Cr(p,γ)53Mn +
53
Cr(p,nγ)53Mn reactions for proton energies ranging between 1.0 and 3 MeV at the laboratory
angle of 900 were measured with a variable energy step (ranging from 20 to 10 keV). The
threshold of the second reaction for 378 keV γ-rays is at 1.79 MeV. The measured γ-ray
production differential cross-sections are shown in Fig. 4.100.
0.25
a
0.20 b
0.15
dσ/dΩ (mb/sr)
0.10
0.05
0.00
800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200
Ep(keV)
FIG. 4.100. a) measured differential cross-section for production of the 378 keV γ-ray from the
52
Cr(p,γ)53Mn + 53Cr(p,nγ)53Mn reactions for proton energies ranging between 900 and 3100 keV at
the laboratory angle of 900, using natCr-target. b) σ TOT/4π for reaction 52Cr(p,γ)53Mn from the work of
[4.124].
The uncertainty budget of the measurement is presented in Table 4.42.

PROTON-INDUCED Cr REACTIONS
Parameter Uncertainty
Target thickness 5%
Integrated beam charges 1%
197
Total cross-sections of 52Cr(p,γ)53Mn have been measured by Kennet et.al. [4.124] for use in
nuclear astrophysics applications. The data were compared with the new measurements in
Fig. 4.100, after they had been converted to differential cross-sections by dividing by a factor
of 4π. This simple conversion works under the assumption that the emitted γ-ray is isotropic.
Gosset used the resonance of the 52Cr(p,γ)53Mn reaction at 1005 keV for profiling Cr that had
been implanted in pure Fe [4.122].
4.14.1.2. Thick target yield on natCr: 378 keV γ-ray
Demortier et al. [4.120] has carried out thick-target yield measurements of the 378 keV γ-ray
on natural chromium. The results are shown in Fig. 4.101.
100000
10000
Yield (Nγ/sr/µC)
1000
100
10
800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200
Ep(keV)
FIG. 4.101. Gamma-ray yield of the 378keV γ-ray as a function of proton energy for natural
Chromium [4.120].
Thick-target yields were calculated from the new measured differential cross-sections. Since
cross-sections at energies below 1000 keV were not measured, it is convenient to compare the
difference of thick-target yields at the current proton beam energy and the thick-target yields
at beam energy 1000 keV (Y(Ep)-Y(1000 keV)). The values of the difference are shown in
Fig. 4.102.
198
5
10
Y(Ep)-Y(1000 keV), (Nγ/(µC*sr))

a
b
4
10
3
10
2
10
1
10
800 1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200
Ep(keV)
FIG. 4.102. The difference in thick target yields at the current proton beam energy and thick target
yields at the beam energy of 1000 keV for the 378 γ-ray. a) thick target yields calculated from the new
measured differential cross-sections at 90°. b) measured thick-target yields at 90° [4.120].

The excitation function for the production of the 378 keV γ-ray produced by the 52Cr(p,γ)53Mn
+ 53Cr(p, nγ)53Mn reactions, as shown in Fig. 4.100, can be used for analytical purposes in the
whole energy range.
[4.1] ANTTILA, A., HÄNNINEN, R., RAISÄNEN, J., Proton-induced thick-target

gamma-ray yilds fort he elemental analysis of the Z = 3-9, 11-21 elements, J.
Radioanal. Nucl. Chem. 62 (1981) 293.
[4.2] KISS, Á., KOLTAY, E., NYAKÓ, B., SOMORJAI, E., ANTTILA, A., RÄISÄNEN,
J., Measurements of relative thick target yields for PIGE analysis on light elements in
te proton energy interval 2.4–4.2 MeV, J. Radioanal. Nucl. Chem. 89 (1985) 123.
[4.3] RÄISÄNEN, J., WITTING, T., KEINONEN, J., Absolute thick-target γ-ray yields
for elemental analysis by 7 and 9 MeV protons, Nucl. Instrum. Methods Phys. Res.
B 28 (1987) 199.
[4.4] RIO, S., MÉTRICH, N., MOSBAH, M., MASSIOT, P., Lithium, boron and
beryllium in volcanic glasses and minerals studied by nuclear microprobe, Nucl.
[4.5] BONI, C., CEREDA, E., BRAGA MARCASSAN, G.M., DE TOMASI, V., Prompt
gamma emission excitation functions for PIGE analysis, Nucl. Instrum. Methods
Phys. Res. B 35 (1988) 80–86.
[4.6] CACIOLLI, A., CHIARI, M., CLIMENT-FONT, A., FERNÁNDEZ-JIMÉNEZ,
M.T., GARCÍA-LÓPEZ, G., LUCARELLI, F., NAVA, S., ZUCCHIATTI, A.,
Measurement of γ-ray emission induced by protons on fluorine and lithium, Nucl.
Instrum. Methods Phys. Res. B 249 (2006) 98–100.
199
[4.7] ZIEGLER, JAMES F., ZIEGLER, M.D., BIERSACK, J.P., SRIM-the stopping and
range of ions in matter (2010), Nucl. Instrum. Methods Phys. Res. B 268 (2010)
1818.
[4.8] EXFOR database [https://www-nds.iaea.org/exfor/exfor.htm].
[4.9] NEWSON, H.W., WILLIAMSON, R.M., JONES, K.W., GIBBONS, J.H., Marshak,
H., Li7(p,n), (p,p'γ), and (p, γ) reactions near neutron threshold, Phys. Rev. 108
(1957) 1294.
[4.10] BACHILLER PAREA, D., CORVISIERO, P., JIMENEZ REY, D., JOCO, D.,
MAIRAL VIDAL, A., MUNOZ MARTIN, A., ZUCCHIATATTI, A., Measurement
of gamma-ray production cross-sections in Li and F induced by protons from 810 to
3700 keV, Nucl. Instrum. Methods Phys. Res. B (2017) in press.
[4.11] MATEUS, R., JESUS, A.P., BRAIZINHA, B., CRUZ, J., PINTO, J.V., RIBEIRO,
J.P., Proton-induced γ-ray analysis of lithium in thick samples, Nucl. Instrum.
[4.12] JESUS, A.P., et al., private communication.
[4.13] BROWN, A.B, SNYDER, C.W., FOWLER, W.A., LARITSEN, C.C., Excited States
of the Mirror Nuclei, Li7 and Be7, Phys. Rev. 82 (1951) 159.
[4.14] PRESSER, G., BASS, R., Reactions 7Li + n, 7Li + p and excited states of the A = 8
system, Nucl. Phys. A 182 (1972) 321.
[4.15] GUZHOVSKIJ, B.J., ABRAMOVICH, S.N., PROTOPOPOV, V.H., Izv. Akad.
Nauk SSSR, Ser. Fiz.-Mat. 4 (1984) 24.
[4.16] ASLAM, W.V., PRESTWICH, MCNEILL, F.E., Thin target 7Li(p,p’g)7 inelastic
gamma-ray yield measurements, J. Radioanal. Nucl. Chem. 254 (2002) 533.
[4.17] SAVIDOU, A., ASLANOGLAU, X., PARADELLIS, T., PILAKOUTA, M., Proton
induced thick target γ-ray yields of light nuclei at the energy region Ep = 1.0-4.1
MeV, Nucl. Instrum. Methods Phys. Res. B 152 (1999) 12–18.
[4.18] ELEKES, Z., KISS, Á.Z., BIRON, I., CALLIGARO, T., SALOMON, J., Thick
target γ-ray yields for light elements measured in the deuteron energy interval of 0.7–
3.4 MeV, Nucl. Instrum. Methods Phys. Res. B 168 (2000) 305.
[4.19] SZIKI, G.Á., SIMON, A., SZIKSZAI, Z., KERTÉSZ, ZS., DOBOS, E., Gamma ray
production cross-sections of deuteron induced nuclear reactios for light element
analysis, Nucl. Instrum. Methods Phys. Res. B 251 (2006) 343.
[4.20] SEPPÄLÄ, A., RÄISÄNEN, J., RAUHALA, E., SZÖKEFALVI-NAGY, Z.,
Absolute thick-target gamma-ray yields for the light elements induced by 12C, 14N
and 16O ions, Nucl. Instrum. Methods Phys. Res. B 143 (1998) 233.
[4.21] PERDIKAKIS, G., SPYROU, A., KOKKORIS, M., ZARKADAS, CH.,
KARYDAS, A.G., HARISSOPULOS, S. KOSSIONIDES, S., On the determination
of beryllium in light element matrices using PIGE and NRA, Nucl. Instrum. Methods
Phys. Res. B 226 (2004) 622.
[4.22] ZAHNOW, D., ANGULO, C., JUNKER, M., ROLFS, C., SCHMIDT, S.,
SCHULTE, W.H., SOMORJAI, E., Nucl. Phys. A 589 (1995) 95.
[4.23] HORNYAK, W. F., LUDEMANN, C. A., ROUSH, M. L., Energy levels of B10,
Nucl. Phys. 50 (1964) 424.
[4.24] FONSECA, M., JESUS, A.P., LUÍS, H., MATEU, R., CRUZ, J., GASQUES, L.,
Instrum. Methods Res. B 268 (2010) 1806.
[4.25] MEYERHOF, W.E., TANNER, N.W., HUDSON, C.M., Radiative capture of
protons by Be9, Phys. Rev 115 (1959) 1227.
[4.26] CALLIGARO, T., DRAN, J.-C., POIROT, J.-P., QUERRÉ, G., SALOMON, J.,
ZWAAN, J.C., PIXE/PIGE characterization of emeralds using an external micro-
beam, Nucl. Inst. Meth. Phys. Res. B 161–163 (2000) 769.
200
[4.27] MA XIN-PEI, MACARTHUR, J.D., ROEDER, P.L., MARIANO, A.N., Trace
element fingerprinting of emeralds by PIXI/PIGE, Nucl. Instrum. Methods Phys.
Res. B 75 (1993) 423.
[4.28] STOTT, P.E., WOOTTON, A., GORINI, G., SINDORNI, E., BATANI, D., (Eds.)
Advanced Diagnostics for Magnetic and Inertial Fusion, 1st edn, Springer US, New
York (2002) p141.
[4.29] BORDERIE, B., BARRANDON, J.N., New analytical developments in prompt
gamma-ray spectrometry with low-energy tritons and alpha particles, Nucl. Instrum.
Methods 156 (1978) 483–492.
[4.30] LAPPALAINEN, R., ANTTILA, A., RÄISÄNEN, J., Absolute α-induced thick-
target gamma-ray hields for the elemental analysis of light elements, Nucl. Instrum.
Methods Phys. Res. 212 (1983) 441.
[4.31] HEATON, P.K., LEE, H. W., ROBERTSON, B. C., NORMAN, E. B., LESKO, K.
T., SUR, B., α-particle induced γ-ray transitions in light elements, Phys. Rev. C 56
(1997) 922.
[4.32] SEGEL R.E., SINGH, P.P., HANNA, S.S., GRACE, M.A., Gamma rays from the
B10+ p: decay schemes and excitation functions, Phys. Rev. 145 (1966) 736–745.
[4.33] DAY, R.B., HUUS, T., Gamma radiation from B10 bombarded by protons, Phys.
Rev. 95 (1954) 1003–1006.
[4.34] LAGOYANNIS, A., PREKETES-SIGALA, K., AXIOTIS, M., FOTEINOU, V.,
HARISSOPULOS, S., KOKKORIS, M., MISAELIDES, P., PANETA, V.,
PATRONIS, N., Study of the 10B(p,αγ)7Be and 10B(p,p’γ)10B reactions for PIGE
purposes, Nucl. Instrum. Methods Phys. Res. B 342 (2015) 271–276.
[4.35] CHIARI, M., FERRACCIOLI, G., MELON, B., NANNINI, A., PEREGO, A.,
SALVESTRINI, L., LAGOYANNIS, A., PREKETES-SIGALAS, K., Measurement
of proton induced thick target γ-ray yields on B, N, Na, Al and Si from 2.5 to 4.1
MeV, Nucl. Instrum. Methods Phys. Res. B 366 (2016) 77.
[4.36] HUNT, S.E., POPE, R.A., EVANS, W.W., Investigation of the gamma radiation
produced by irradiating B10 with protons in the energy range 0.7 to 3.0 MeV, Phys.
Rev. 106 (1957) 1012–1015.
[4.37] OPHEL, T.R., GLOVER, R.N., TITTERTON, E.W., Gamma ray yields from the
proton bombardment of B10, Nucl. Phys. 33 (1962) 198 – 201.
[4.38] MATEUS, R., REIS, M.A., JESUS, A.P., RIBEIRO, J.P., Quantitative analysis of
light elements in aerosol samples by PIGE, Nucl. Instrum. Methods Phys. Res. B
249 (2006) 784–788.
[4.39] PREKETES-SIGALAS, K., LAGOYANNIS, A., AXIOTIS, M., HARISSOPULOS,
S., KOKKORIS, M., MERTZIMEKIS, T.J., PANETA, V., PROVATAS, G., Study
of the 11B(p,p′γ)11B reaction for PIGE applications, Nucl. Instrum. Methods Phys.
Res. B 368 (2016) 71–74.
[4.40] GURBICH, A. F., ISTOMIN, I. V., SEMENOV, A. V., A nuclear-physics method of
composition analysis for the wall surface of a BN-600 steam generator, Sov. J. Atom.
Energy 68 (1990), 236–239.
[4.41] ZHANG WEICHENG, LEI XIANGGUO, Proton-induced gamma-ray emission for
carbon tracer techniques, Appl. Radiat. Isot. 42 (1991) 1039.
[4.42] ZHANG WEICHENG, ZHANG TIANMEI, LEI XIANGGUO, Determination of 13C
in breath sample by PIGE, J. Radioanal. Nucl. Chem. 151 (1991) 313.
[4.43] ROLFS, C., AZUMA, R. E., Interference effects in 12C(p,γ)13N and direct capture to
unbound states, Nucl. Phys. A 227 (1974) 291.
[4.44] BURTEBAEV, N., IGAMOV, S.B., PETERSON, R.J., YARMUKHAMEDOV, R.,
ZAZULIN, D.M., New measurements of the astrophysical S factor for 12C(p,γ)13N
201
reaction at low energies and the asymptotic normalization coefficient (nuclear vertex
constant) for the p + 12C 13N reaction, Phys. Rev. C 78 (2008) 035802.
[4.45] GOERRES, J., private communication.
[4.46] KING, J.D., AZUMA, R.E., VISE, J.B., Cross section and astrophysical S-factor for
the 13C(p,γ)14N reaction, Nucl. Phys A 567 (1994) 345.
[4.47] BAKTIBAYEV, M.K.,BURMINSKII, V.P., BURTEBAYEV, N., JAZAIROV-
KAKHRAMONOV, V., KADYRZHANOV, K.K., SAGINDYKOV, SH.SH.,
ZARIFOV, R.A., ZAZULIN, D.M., Differential Cross Section Measurement of
Radiative Capture of Protons by Nuclei 13C (Proc. 3rd Eurasian Conf. Nucl. Sci.
Appl. Uzbekistan, 2004) Vol. 38, INIS-UZ--123, Uzbekistan (2004) 375.
[4.48] STANGEBY, P.C., et al., C-Tracer Experiments in DIII-D Preliminary to Thermal
Oxidation Experiments to Understand Tritium Recovery in DIII-D, JET, C-Mod, and
MAST, (Proc. 33rd EPS Conf. on Plasma Phys. Rome, 2006) ECA Vol. 30I,
European Physical Society, Mulhouse (2006) 5,145.
[4.49] CABIOC’H, T., THUNE, E., JAOUEN, M., BODART, F., Resonant nuclear
reaction analysis study of carbon phase formation during carbon ion-implantation
itno silver substrates, Nucl. Instrum. Methods Phys. Res. B 207 (2003) 409.
[4.50] THEODOSSIU, W., BAUMANN, H., MARKWITZ, A., BETHGE, K.,
Characterisation of 13C implantations in silicon by NRA [13C(p,γ)14N] and RBS,
Fresenius’ J. Anal. Chem. 353 (1995) 483.
[4.51] KIENER, J., GROS, M., TATISCHEFF, V., ATTIÉ, D., BAILLY, I., BAUCHET,
A., CHAPUIS, C., CORDIER, B., DELONCLE, I., PORQUET, M.G., SCHANNE,
S., SÉRÉVILLE, N.DE, TAUZIN, G., New determinations of gamma-ray line
intensities of the Ep + 550 and 1747 keV resonances of the 13C(p,γ)14N reaction,
Nucl. Instrum. Methods Phys. Res. A 519 (2004) 623.
[4.52] DYER, P., BODANSKY, D., SEAMSTER, A.G., NORMAN, E.B., MAXSON,
D.R., Cross sections relevant to gamma-ray astronomy: proton induced reactions,
Phys. Rev. C 23 (1981) 1865.
[4.53] PIERCE, T.B., PECK, P.F., HENRY, W.M., Determination of carbon in steels by
measurement of the prompt γ-radiation emitted during proton bombardment, Nature
4958 (1964) 571.
[4.54] ENE, A., POPESCU, V., BADICA, T., Determination of carbon in steel using
particle-induced gamma ray spectrometry, J. Optoelectron. Adv. M. 8, (2006) 222–
224.
[4.55] LAPATTO, R., RÄISÄNEN, J., A rapid method for the quantification of C,N and O
in biomedical samples by proton-induced gamma-ray emission analysis applied to
human placental samples, Phys. Med. Biol. 33 (1988) 75.
[4.56] BAIR, J.K., COHN, H.O., WILLARD, H.B., Level structure of N14, Phys. Rev. 119
(1960) 2026.
[4.57] BANERJEE, M., SRINIVASAN, N.B., ZHU, H., KIM, S.J., XU, K., WINTER, M.,
BECKER, H-W., ROGALLA, D., DE LOS ARCOS, T., BEKERMANN, D.,
BARRECAI, D., FISCHER, R.A., DEVI, A., Fabrication of ZrO2 and ZrN films by
metalorganic chemical vapor deposition employing new Zr precursors, Cryst.
Growth Des. 12 (2012) 5079.
[4.58] BEKERMANN, D., LUDWIG, A., TOADER, T., MACCATO, C., BARRECA, D.,
GASPAROTTO, A., BOCK, C., WIECK, A.D., KUNZE, U., TONDELLO, E.,
FISCHER, R.A., DEVI, A., MOCVD of ZnO films from Bis(ketoiminato)Zn(II)
precursors: structure, morphology and optical properties, Chem. Vap. Deposition 17
(2011) 155.
[4.59] CSEDREKI, L., UZONYI, I., SZÍKI, G.Á., SZIKSZAI, Z., GYÜRKY, GY., KISS,
Á.Z., Measurements and assessment of 12C(d,pγ)13C reaction cross sections in the
202
deuteron energy range 740–2000 keV for analytical applications, Nucl. Instrum.
[4.60] TRYTI, S., HOLTEBEKK, T., REKSTAD, J., Angular distributions of protons near
resonance for the reaction 12C(d,pγ)13C obtained by shape studies of γ-ray lines,
Nucl. Phys. A 201 (1973) 135–144.
[4.61] TRYTI, S., HOLTEBEKK, T., UGLETVEIT, F., Angular distributions of protons
from the reaction 12C(d,pγ)13C obtained by shape studies of γ-ray lines, Nucl. Phys.
A 251 (1975) 206–224.
[4.62] PAPILLON, F., WALTER, P., Analytical use of the multiple gamma-rays from the
12
C(d,pγ)13C* nuclear reaction, Nucl. Instrum. Methods Phys. Res. B 132 (1997)
468–480.
[4.63] MARCHAND, B., MIZOHATA, K., RÄISÄNEN, J., Proton induced gamma-ray
production cross sections and thick-target yields for boron, nitrogen and silicon,
Nucl. Instrum. Methods Phys. Res. B 378 (2016) 25–30.
[4.64] AJZENBERG-SELOVE, F., Energy levels of light nuclei A = 13–15, Nucl. Phys. A
523 (1991) 1–196.
[4.65] RÄISÄNEN, J., TIKKANEN, P.O., Calibration of accelerator proton energy by the
14
N(p,p′γ)14N reation resonances, Nucl. Instrum. Methods Phys. Res. A 723 (2013)
5.
[4.66] PHILLIPS, G.W., RICHARD, P., ELLIOT, D.O., HOPKINS, F.F., PORTER, A.C.,
Structure of mass 15: the 14N(p,γ)15O and 14N(p,p′γ) reactions, Phys. Rev. C 5 (1972)
297.
[4.67] BENHABILES-MEZHOUD, H., KIENER, J., THIBAUD, J.-P., TATISCHEFF, V.,
DELONCLE, I., COC, A., DUPRAT, J., HAMADACHE, C., LEFEBVRE-
SCHUHL, DALOUZY, J.-C., DE GRACEY, F., DE OLIVEIRA, F., A., DAYRAS,
DE SÉRÉVILLE, N., PELLEGRITI, M.-G., F., LAMIA, L., OUICHAOUI, S.,
Measurements of nuclear γ-ray line emission in interactions of protons and α
particles with N, O, Ne, and Si, Phys. Rev C 83 (2011) 024603.
[4.68] WEST, M.L., JONES, C.M., BAIR, J.K., WILLARD, H.B., Level structure in 15O
from the proton bombardment of 14N, Phys. Rev. 179 (1969) 1047.
SALVESTRINI, L., LAGOYANNIS, A., PREKETES-SIGALAS, K., Measurment
MeV, Nucl. Instrum. Methods Phys. Res. B 366 (2016) 77; Ion Beam Analysis
Nuclear Data Library (IBANDL), http://www-nds.iaea.org/ibandl/
[4.70] GORODETZKY, S., ADLOFF, J.C., BROCHARD, F., CHEVALLIER, P.,
DISPIER, D., GORODETZKY, PH., MODJTAHED-ZADEH, R., SCHEIBLING,
F., Cascades γ-γ dee quatre rÉsonances de la réaction 15N(p,γ)16O, Nucl. Phys. A 113
(1968) 221.
[4.71] ROLFS, C., RODNEY, W.S., Proton capture by 15N at stellar energies, Nucl. Phys.
A 235 (1974) 450.
[4.72] IMBRIANI, G., DEBOER, R. J., BEST, A., COUDER, M., GERVINO, G.,
GÖRRES, J., LEBLANC, P. J., LEISTE, H., LEMUT, A., STECH, E., STRIEDER,
F., UBERSEDER, E., WIESCHER, M., Measurement of γ rays from 15N(p,γ)16O
cascade and 15N(p,α1γ)12C reactions, Phys Rev. C 85 (2012) 065810.
[4.73] HÄNNINEN, R., RÄISÄNEN, J., Nitrogen determination with external beam proton
induced gamma-ray emission analysis, Nucl. Instrum. Methods Phys. Res. B 5
(1984) 34–38.
203
[4.74] VAN BEBBER, H., BORUCKI, L., FARZIN, K., KISS, Á.Z., SCHULTE, W.H.,
Total cross section of the 14N(d,pγ)15N nuclear reaction for analytical applications,
Nucl. Instrum. Methods Phys. Res. B 136–138 (1998) 72.
[4.75] CSEDREKI, L., UZONYI, I., SZIKSZAI, Z., GYÜRKY, GY., SZÍKI, G.Á., KISS,
Á.Z., Measurement of gamm-ray production cross section for nuclear reactions
14
N(d,pγ)15N and 28Si(d,pγ)29Si, Nucl. Instrum. Methods Phys. Res. B 328 (2014)
20–26.
[4.76] CSEDREKI, L., Experimental conditions for cross-section measurements for
analytical purposes, Acta Phys. Debrecina 46 (2012) 25–35.
[4.77] CSEDREKI, L., HALÁSZ, Z., KISS, Á.Z., Assessment of experimental d-PIGE γ-
ray production cross-sections for 12C, 14N and 16O and comparison with absolute
thick target yields, Nucl. Instrum. Methods Phys. Res. B 380 (2016) 1–10.
[4.78] Unpublished data found on IBANDL [https://www-nds.iaea.org/exfor/ibandl.htm].
[4.79] ZAMBONI, I., SIKETIĆ, Z., JAKŠIĆ, M., BOGDANOVIĆ RADOVIĆ, I.,
Measurement of proton induced γ-ray emission cross-sections on MgF2 target in the
energy range 1.95–3.05 MeV, Nucl. Instrum. Methods Phys. Res. B 342 (2015)
266–270.
[4.80] ALLEGRO, P.R.P., DE ALMEIDA RIZZUTTO, M., MEDINA, N.H.,
Improvements in the PIGE technique via gamma-ray angular distribution,
Microchem. J. 126 (2016) 287–295.
[4.81] CHAUDRI, M. A., The absolute total cross-section and thick target yields of the
19
F(p,p′γ)19F reaction from 0.4 to 4.3 MeV, Radiat. Eff. Defect. S. 94 (1986) 145–
150.
[4.82] JESUS, A.P., BRAIZINHA, B., RIBEIRO, J.P., Excitation function and cross-
section of the reaction 19F(p,p′γ) 19F, Nucl. Instrum. Methods Phys. Res. B 161–163
(2000) 186–190.
[4.83] GRAMBOLE, D., BAUER, C., GIPPNER, P., PHEISER, C., RUDOLPH, W.,
THOMAS, H.J., Fluorine determination in the near-surface region of solids using the
19
F(p,p'γ)19F resonance reaction, J. Radioanal. Nucl. Chem. 83 (1984) 107.
[4.84] RANKEN, W.A., BONNER, T.W., MCCRARY, J.H., Energy dependence of F19 + p
reactions, Phys. Rev. 109 (1958) 1646–1651.
[4.85] LAGOYANNIS, A., private communication.
[4.86] KENNY, M.J., Thick target gamma yields from fluorine, Austr. J. Phys. 34 (1981)
35–42.
[4.87] MATEUS, R., JESUS, A.P., RIBEIRO, J.P., Quantitative analysis of light elements
in thick samples by PIGE, Nucl. Instrum. Methods Phys. Res. B 219–220 (2004)
519–523.
[4.88] CHIARI, M., ALVES, E. BOGDANOVIĆ RADOVIĆ, I., CRUZ, J., CSEDREKI, L.
FONSECA, M., GALAVIZ, D., HENRIQUES, A., JAKŠIC, M., JESUS, A.P.,
KAKUEE, O., KISS, Á.Z., LAGOYANNIS, A., LOURENÇO, F., LUÍS, H.,
MACHADO, J., MELON, B., NUVIADENU, C.K., SALVESTRINI, L.,
SHARIFZEDEH, N., SIKETIĆ, Z., SZÍKI, G.Á., SZIKSZAI, Z., TEUBIG, P., VELHO,
P., ZAMBONI, I., ZARZA, M., Measurement of proton induced γ-ray emission
cross-sections on Na from 1.0 to 4.1 MeV, Nucl. Instrum. Methods Phys. Res. B
(2017) available online.
[4.89] CACIOLLI, A., CALZOLAI, G., CHIARI, M., CLIMENT-FONT, A., GARCIA, G.,
LUCARELLI, F., NAVA, S., Proton elastic scattering and proton induced γ-ray
emission cross-sections on Na from 2 to 5 MeV, Nucl. Instrum. Methods Phys. Res.
B 266 (2008) 1392.
204
[4.90] MATEUS, R., JESUS, A.P., CRUZ, J., RIBEIRO, J.P., Measurement of the inelastic
scattering of protons by 23Na in the energy range 1.25–2.40 MeV, Nucl. Instrum.
Methods Phys. Res. B 219–220 (2004) 307.
[4.91] TROCELLIER, P., ENGELMANN, CH., Fonctions d’excitation des reaction
snucleaires:23Na(p,p′γ)23Na, 23Na(p,αγ)20Ne et23Na(p,α0)20Ne entre 500 et 1700 keV,
J. Radioanal. Nucl. Chem. 67 (1981) 135–142.
[4.92] CALZOLAI, G., CHIARI, M., LUCARELLI, F., NAVA, S., PORTARENA, S.,
Proton induced γ-ray emission yields for the analysis of light elments in aerosol
samples in an external beam set-up, Nucl. Instrum. Methods Phys Res. B 268 (2010)
1540.
[4.93] STELSON, P.H., PRESTON, W.M., Resonant states of Mg24 excited by protons on
sodium, Phys. Rev. 95 (1954) 974.
[4.94] HIRATATE, Y., TAKEUCHI, Y., YAMAYA, T., MORITA, S., Some studies of the
23
Na + p reaction at Ep = 2.2–4.5 MeV, J. Phys. Soc. Jpn. 24 (1968) 690.
[4.95] ASKING, L., SWIETLICKI, E., GARG, M.L., PIGE analysis of sodium in thin
aerosol samples, Nucl. Instrum. Methods Phys. Res. B 22 (1987) 368.
[4.96] BODART, F., DECONNINCK, G., DEMORTIER, G., Quantitative analysis of
sodium by (p,γ) reactions, J. Radioanal. Nucl. Chem. 35 (1977) 95.
[4.97] GURBICH, A.F., Evaluation of cross-sections for particle induced gamma-ray
emission (PIGE) spectroscopy, Nucl. Instrum. Methods Phys. Res. B 331 (2014) 31.
[4.98] SALAH, H., TOUCHRIFT, B., Multielemental analysis of zeolites using prompt
nuclear reactions gamma rays, Nucl. Instrum. Methods Phys. Res. B 129 (1997)
261–265.
[4.99] SMIT, Z., JEZERŠEK, D., KNIFIC, T., ISTENIĆ, J., PIXE-PIGE analysis of
Carolingian period glass from Slovenia, Nucl. Instrum. Methods Phys. Res. B. 267
(2009) 121–124.
[4.100] SHARIFZADEH, N., KAKUEE, O., MOHAMMADI, S., Measurements of proton
induced γ-ray emission cross-sections on Mg from 1.0 to 3.0 MeV, Nucl. Instrum.
[4.101] CHIARI, M., MELON, B., SALVESTRINI, L., FONSECA, M., ALVES, E., JESUS,
A.P., Measurement of proton induced γ-ray emission cross-sections on Al from 2.5
to 4.1 MeV, Nucl. Instrum. Methods Phys. Res. B 332 (2014) 355–358.
[4.102] JOKAR, A., KAKUEE, O., LAMEHI-RACHTI, M., SHARIFZADEH, N.,
FATHOLLAHI, V., Differential cross-section measurements of 27Al (p,p′γ) 27Al and
27
Al(p,αγ)24Mg reactions in the energy range of 1.6–3.0 MeV, Nucl. Instrum.
[4.103] JOKAR, A., KAKUEE, O., LAMEHI-RACHTI, M., FATHOLLAHI, V., Thick
target yields of proton induced gamma-ray emission from Al, Si and P, Nucl. Instr.
Phys. Meth. Res. B 394 (2017) 28–32.
[4.104] WARSH, K.L., TEMMER, G.M., BLIEDEN, H.R., The Al27(p,α)Mg24 reaction with
3 to 12 MeV protons, Nucl. Phys. 44 (1963) 329–337.
ROLFS, C., SCHULTE, W.H., BECKER, H.W., The 27Al(p, γ)28Si reaction: direct
capture cross-section and resonance strengths at Ep = 0.2–1.12 MeV, Eur. Phys. J. A
9 (2000) 479.
[4.106] WANG, Y., NASTASI, M., (eds), Handbook of Modern Ion Beam Materials
Analysis, 2nd edn, Materials Research Society, Warrendale, Pennsylvania (2010).
[4.107] DECONNINCK, G., DEMORTIER, G., Quantitative analysis of aluminum by
prompt nuclear reactions, J. Radioanal. Nucl. Chem. 12 (1972) 189–208.
205
[4.108] JOKAR, A., KAKUEE, O., LAMEHI-RACHTI, M., Differential cross-sections
measurement of 28Si(p, p′γ)28Si and 29Si(p,p′γ)29Si reactions for PIGE applictions,
[4.109] RÄISÄNEN, J., HIRVONEN, J., ANTTILA, A., The diffusion of silicon in
germanium, Solid-State Electron. 24 (1981) 333-336.
[4.110] RÄISÄNEN, J., KEINONEN, J., Annealing behaviour of Si in ion-implanted α-Ti,
Appl. Phys. Lett. 49 13 (1986) 773–775.
[4.111] AHLGREN, T., LIKONEN, J., SLOTTE, J., RÄISÄNEN, J., RAJATORA, M.,
KEINONEN, J., Concentration dependent and independent Si diffusion in ion-
implanted GaAs, Phys. Rev. B 56 (1997) 4597–4603.
SALVESTRINI, L., LAGOYANNIS, A., PREKETES-SIGALAS, K., Measurement
MeV, Nucl. Instrum. Methods Phys. Res. B 366 (2016) 77–82.
[4.113] CSEDREKI, L., UZONYI, I., SZIKSZAI, Z., GYÜRKY, GY., SZÍKI, G.Á., KISS,
Á.Z., Measurement of gamma-ray production cross-sections for nuclear reactions
14N(d,pγ)15N and 28Si(d,pγ)29Si, Nucl. Instrum. Methods Phys. Res. B 328 (2014)
20–26.
[4.114] JOKAR, A. KAKUEE, O., LAMEHI-RACHTI, M., Differential cross-section
smeasurement of 31P(p,pγ1)31P reaction for PIGE applications, Nucl. Intrum.
Methods Phys. Res. B 383 (2016) 152-155.
[4.115] DEMORTIER, G., BODART, F., Quantitative analysis of phosphorus by (p,γ)
reactions, J. Radioanal. Nucl. Chem. 12 (1972) 209–219.
[4.116] LAGOYANNIS, A., et al., private communication.
[4.117] RAO, P., KUMAR, S., VIKRAMKUMAR, S., RAJU, V.S., Measurement of
differential cross-sections and widths of resonances in 32S(p,p′γ)32S reaction in the
3.0–4.0 MeV region, Nucl. Instrum. Methods Phys. Res. B 269 (2011) 2557
[4.118] TSARTSARAKOS, C., MISAELIDES, P., KATSANOS, A., Cross-sections for the
32
S(p,p′γ)32S nuclear reaction used for the profiling of sulphur on materials surfaces,
[4.119] OLNESS, J. W., HAEBERLI, W., LEWIS, H.W., Levels of Cl33 from S32(p,p)S32
and S32(p,p′γ)S32, Phys. Rev. 112 (1958) 1702.
[4.120] DEMORTIER, G., Prompt gamma-ray yields from proton bombardment of transition
elmeents (Ti to Zn), J. Radioanal. Nucl. Chem. 45 (1978) 459–496.
[4.121] KENNET, S.R., MITCHELL, L.W., ANDERSEN, M.R., SARGOOD, D.G., Cross-
section measurements and thermonuclear reaction rates for 46Ti(p,γ)47V, 47Ti(p,γ)48V,
47
Ti(p,n)47V and 48Ti(p, γ)49V, Nucl. Phys. A 368 (1981) 337–351.
[4.122] GOSSETT, C.R., A method for determining depth profiles of transition elements in
steels, Nucl. Instrum. Methods 168 (1980) 217–221.
[4.123] KRIVONOSOV, G.A., NEMASHKALO, B.A., STORIZHKO, V.E.,
KLYUCHAREV, A.P., EKHICHEV, O.I., CHIRT, V.K., Inelastic scattering of
protons by medium-weight nuclei at low energies, Yad. Fiz. 24 (1976) 461.
[4.124] KENNET, S.R., MITCHELL, L.W., ANDERSON, M.R., SARGOOD, D.G., Cross-
section measurements and thermonuclear reaction rates for 52Cr(p,γ)53Mn and
54
Fe(p,γ)55Co, Nucl. Phys. A 363 (1981) 233.
206
5. THE EMITTED RADIATION YIELD ANALYSIS (ERYA) CODE FOR PIGE
ANALYSIS
5.1. INTRODUCTION
Emitted Radiation Yield Analysis (ERYA) [5.1] is a code developed for PIGE analysis. It
uses relevant quantities and experimental parameters to calculate the concentrations of
elements/isotopes, given their gamma-ray yield (and vice-versa). At the onset, ERYA was
programmed only for in-depth homogeneous samples, i.e. bulk analysis. In order to deal with
depth-dependent concentrations, ERYA-profiling was developed, which is a code based on
ERYA. Both codes are discussed in more detail in the following sections.
5.2. ERYA
5.2.1. ERYA-bulk
Section 2.1 provides the basis of the calculations made by ERYA [5.2]. Eq. 5.1 calculates the
yield of gamma-rays for an in-depth homogeneous sample, with energy Eγ detected at a given
ion beam energy E0. The yield is assumed to be isotropic.
Q E0
Y ( E ) = ε abs ( E γ ) ⋅   ⋅ f m ⋅ f i ⋅ N av ⋅ A −1 ⋅ ∫ σ ( E ) S m ( E ) dE, (5.1)
e 0
where Q/e is the number of bombarding ions, Q is the collected charge, e is the electric charge
of the ion beam, σ is the relevant nuclear reaction cross-section, εabs is the absolute efficiency
of detection, fi is the isotopic abundance, Nav is the Avogrado number, A is the atomic mass
of the relevant element and Sm(E) is the mass stopping power, expressed in energy per areal
mass units.
In order to perform the calculation, specifically the integration, ERYA divides the sample into
n sublayers that are perpendicular to the incident beam, i.e. the energy range of the incident
ions is divided into several energy intervals. The magnitude of each interval is such that,
within each interval, the stopping power cross-section can be assumed to be constant. The
energy intervals may be set equal to the energy steps that are necessary to accurately define
the employed excitation function. This means that the energy steps are small enough to
assume that, in between any two consecutive values of the energy, the excitation function has
a linear dependence on the energy. Taking the above considerations into account, the integral
may be approximated by the following sum:
∫
E0
0
σ(E)/ε(E) dE =∑ (1 S mk ) ⋅ ∫


Ek + 1
Ek  k
[ ]
σ(E) dE  = ∑ (1 S mk ) ⋅ σ(E k,k + 1 ) ⋅ ( E k +1 − E k ) , (5.2)
k
where the cross-section for each layer k is the mean between σk and σk + 1, i.e. the measured
cross-sections at energies Ek and Ek + 1. The integrated cross-section increases rapidly with the
proton energy until the minimum energy in the cross-section data file Emin is reached, which
replaces zero, in this case, as the end point of integration.
207
The code was developed in a Labview-based interface in order to be user-friendly. The main
interface is shown in Fig. 5.1. The eight inputs listed below were introduced via this interface
and are necessary to perform the proper calculations.
(i) Efficiency of the gamma-ray detector — accessible through the “database” button on
the main interface (Fig. 5.1). Users can also upload a text file to the code folder, with a
table of efficiency values for a given energy range or function of the gamma energy. A
specific function or number for a particular gamma-ray may be entered via the main
interface (Fig. 5.1). Number of elements and choice of elements may also be entered via
the main interface (Fig. 5.1) and includes the elements/isotopes to be quantified by
PIGE, which may be minor or trace components of the sample, and the elements which
are major components of the sample composition. This information is important for
calculation of the stopping power. There is a database associated with ERYA that
contains the atomic number and masses of some elements and respective isotopic
abundances (Fig. 5.2). This information will be automatically uploaded when the user
chooses an element, but the user may also access the database through a button on the
main interface and modify or add information as needed (Fig. 5.2).
(ii) Cross-Section of the gamma-producing reaction used to quantify the

corresponding isotope — accessible through the “database” button on the main
interface (Fig. 5.1). There is a database of excitation functions, which were measured or
adopted by the Lisbon group and that is connected with ERYA. This database reduces
input from individual users since the appropriate isotope can just be selected via the
interface. If more than one gamma-ray is associated with the isotope, the user must
choose the gamma-ray as well. The user may also replace the existing excitation
functions through the same menu as referred to in #2. Direct input from IBANDL
database is also possible.
(iii) Stopping power — accessible through the “database” button on the main interface (Fig.
5.1). The user will find a table of Ziegler parameters (see the respective window, in an
inset over the graph in Fig. 5.1) for the semi-empirical calculation of the stopping
power. The user can modify these parameters as needed.
(iv) First guess of the sample composition — atomic fractions of the sample components,
which must sum up to 1, are entered in the “%”column (Fig. 5.1).
(v) Gamma-ray experimental yields — entered in the column marked “YoExp” (Fig.
5.1). These must correspond to 1 µC of collected charge, which is the default value. If
not, the user must enter the charge in the respective field on the main interface (Fig.
5.1).
(vi) Target thickness — entered in the “thick” field (Fig. 5.1). A value only needs to be
added for a thin or semi-thick sample.
(vii) Proton energy of analysis — entered in the “energy max” field on the main interface
(Fig. 5.1). The sum in Eq. 5.2 will start from this ‘maximum’ energy and will proceed
by summing terms in well-defined energy steps until the minimum energy in the cross-
section data file Emin is reached. A different step than the experimental step used in the
measurement may be chosen from the “step size” field on the main interface (Fig. 5.1).
The default value is 1 keV.
208
(viii) Calibration parameter — entered in the “CP” column on the main interface (Fig. 5.1).
The calibration parameter (see Section 2.1) may be inserted in order to take into
consideration the experimental system of the user.
After (i–viii) have been entered, the ERYA main interface will display the resulting yields
that correspond to the inserted atomic fractions in the column marked “YoSimul”. Users can
then compare “YoSimul” with “YoExp” and modify the introduced atomic fractions until
“YoSimul” agrees with “YoExp”.
The ERYA code can also perform a fitting operation, which employs a non-linear Levenberg-
Marquar algorithm [5.3, 5.4] and provides the optimal sample composition by simultaneously
fitting the gamma-ray yields of all the elements in the sample to be quantified. The user can
choose either of the following options:
(a) Set the initial guess for some of the elements and allow the fit to modify the others (see
the corresponding column on the main interface and the example in Fig. 5.1);
(b) Group the elements in order to keep their atomic ratio constant and allow the fit to
modify the specific amounts of each element (fitting mode field on the main interface in
Fig. 5.1).
The information provided for the major component of the sample composition usually has
two sources: (i) other analytical sources, e.g. from PIXE analysis and (ii) the chemical
content. This results in the following situations, which explain the corresponding two options
above:
(a) From PIXE analyses one obtains concentrations of elements heavier than Na, in the case
of thin window detectors, and heavier than Mg, in all other cases. In order to reach a
concentration sum of 100%, the undetected lighter elements are usually assumed to be
oxygen, which is the main light element of most samples. In this case, the ERYA fit
should use the PIXE data as a first guess and the fit should be performed on the
elemental concentrations of the elements seen by PIGE (even if also detected by PIXE)
and the oxygen concentration. The concentrations of the other elements, which are only
detected by PIXE, should remain fixed;
(b) If the major chemical composition of the sample is known, assuming that the elements
to be quantified by PIGE are only minor or trace elements of the sample, the major
chemical composition may be given as the first guess. The fit should be performed with
the corresponding major elements grouped so that the ratios will stay constant while the
sum of the concentrations is allowed to vary.
In the example given in Fig. 5.1, given option (b), Ag would be included in the fitting
procedure, and Li and F would be grouped to keep the LiF stoichiometry. As a result, a better
fit would be achieved.
To conclude this section on ERYA-bulk, the flow chart of the main code is illustrated in Fig.
5.3.
209
FIG. 5.1. Main interface of ERYA. The meaning of each different field is explained in the text. In this case, a mixture of Ag powder with LiF is displayed before
the fitting procedure. The first guess of the atomic fraction was entered under the “%”column. In the inset (blue square), results are shown from after the
fitting, where “% fit” and “% mass” give the atomic fractions and the mass fractions respectively.
210
FIG. 5.2. The database associated with ERYA. The scale of the graph may be modified in order to zoom into a specific energy range. Information about
new elements/isotopes, including different excitation functions, can be inserted through the “add element” window.
211
FIG. 5.3. Flow chart of the main code, highlighting the choices and parameters to be introduced by the user.
212
5.2.2. ERYA-profiling
Resonant nuclear reaction depth profiling (see Section 2.2) is based on the following
assumption: if a proton beam with energy equal to ER, where ER is the energy of a resonance
of the nucleus in the sample, bombards a sample, only the nuclei at the surface contribute to
the gamma-ray yield. Since the nuclei contributing to the gamma-ray yield are the ones
corresponding to energy ER, increasing the incident energy moves the resonance inside the
sample. The incoming protons shed part of their incident energy at a certain depth, which
depends on the incident energy and the stopping power of protons in the sample material.
Therefore, by increasing the energy above ER, gamma-ray yield can be measured as a function
of the incident energy, which, for thin isolated resonances, gives the concentration of the
given element/isotope of the sample with respect to the depth in the sample.
In reality, the obtained profile is influenced by the energy resolution of the proton beam and
the energy straggling of the proton beam along the depth of the sample. Also, if more than
one resonance contributes to the yield, certain calculations need to be performed in order to
obtain the correct depth profile of the sample. ERYA-profiling is a code that has been
developed to perform these calculations by using a database of excitation functions in
numeric form (the same as for ERYA-bulk) and taking into consideration the energy
resolution of the proton beam and the energy straggling of the proton beam along the depth of
the sample.
Regarding energy straggling, different approximations are implemented, i.e. the

approximations of Landau, Vavilov and Gaussian [5.5–5.7]. These three approximations are
considered to be valid for the amount of energy lost in the samples. To this date, this is the
first code to implement energy straggling to such a high degree of accuracy. For in-depth
homogeneous samples, the gamma-ray yield is given by:
Q E0
Y ( E ) = εabs ( Eγ ) ⋅   ⋅ f m ⋅ f i ⋅ N av ⋅ A−1 ⋅ ∫ σ ( E ) S m ( E ) dE (5.3)
e 0
For in-depth heterogeneous samples, i.e. with mass fraction dependent on depth, which is
equivalent to being dependent on energy, Eq. 5.3 can be modified to Eq. 5.4:
Q E0
Y ( E ) = εabs ( Eγ ) ⋅   ⋅ f i ⋅ N av ⋅ A−1 ⋅ ∫ f m ( E ) ⋅ σ ( E ) S m ( E ) dE (5.4)
e 0
If, for each layer contributing to the integral, the protons do not have the same energy, but
rather follow an energy distribution FE (E ) with mean proton energy E , the equation above
can be transformed into:
Q E0 f (E ) ∞
Y ( E ) = εabs ( Eγ ) ⋅   ⋅ f i ⋅ N av ⋅ A−1 ⋅ ∫ m
⋅  ∫ FE ( E ) ⋅ σ ( E ) ⋅ dE  ⋅ dE (5.5)
0 S (E )  0 
e m
If the sample is divided into n layers thin enough so that the variations of the mean energy of
the stopping power and of the mass fraction are negligible, the yield can be given by:
k =n
Q f (E ) ∞
Y ( E ) = εabs ( Eγ ) ⋅   ⋅ f i ⋅ N av ⋅ A−1 ⋅ ∑ m k ⋅  ∫ FE ( E ) ⋅ σ ( E ) ⋅ dE  ⋅ ∆Ek (5.6)
e 0
k =1 S m ( Ek ) 

213
where the integral within brackets can also be calculated as a sum of a finite small number of
energy values.
ERYA-profiling uses the same database as ERYA-bulk for excitation functions σ(E) and
stopping powers Sm(E). Therefore, it was a natural choice to implement the code in the same
Labview platform. The computer program C was used for the more involved distribution
calculations.
The implemented Labview interface is user-friendly, with several fields for entering
information, including (Fig. 5.4):
(a) Number of elements necessary for profiling or calculating stopping power (major
composition of the sample)
(b) Type of elements
(c) Gamma-ray chosen for profiling
(d) Collected beam charge (default is 1 µC)
(e) Minimum incident energy
(f) Maximum incident energy
(g) Concentration of elements versus depth (atoms per surface unit). Alternatively, the user
can specify the density, linear thickness and composition (stoichiometry) of each layer.
(h) Beam energy resolution
(i) Sample temperature (a Doppler broadening distribution is also implemented in the code)
This code applies a non-automatic, user-based, iterative approach to obtain the concentration
of the element versus depth. Users enter an initial guess of the concentration and compare the
experimental yields obtained with the calculated ones. Next, they modify the guess in order to
improve the match between measured and calculated yields.
214
FIG. 5.4. ERYA-profiling main interface. The meaning and purpose of all buttons and fields are described in the text.
215
The interface has a window to upload the experimental yield versus incident energy data. The
calculated yield appears in the same window, after the calculation is complete (Fig. 5.5).
Several options are available to specify the concentration of the desired element at various
depths: (a) directly through the interface or (b) uploading a text file with either number of
atoms of the element versus number of atoms per area of the sample, or with number of atoms
per area of the element versus beam energy.
In order to more accurately define the resonance to be used for analysis, the user may choose
to use a Breit-Wigner function between Emin and Emax, as long as the resonance energy ER, the
width of the resonance ΓR, and the maximum cross-section value are known. The values of
Emin and Emax may be chosen by the user. Default values used by the code correspond to ER ±
5xΓR.
Figure 5.6 and 5.7 show flowcharts corresponding to the main routines.
216
FIG. 5.5. Interface page for the comparison between experimental and calculated yields as a function of incident energy.
217
FIG. 5.6. Flowchart of ERYA-profiling main routine, highlighting the choices and parameters to be introduced by the user.
218
FIG. 5.7. Flowchart of the AllLayers routine, which is visible in the lower right corner of the flowchart of Fig. 5.6.
219
5.3. USER FEEDBACK AND FUTURE DEVELOPMENTS
Since 2013, the ERYA-bulk code has been available to be downloaded from the following
webpage: http://www.ctn.tecnico.ulisboa.pt/sec/fis/rn/uk_rn_erya.htm (Fig. 5.8), where a
comprehensive user’s manual is also available.
FIG. 5.8. Screenshot of the ERYA webpage.
Feedback from users has generally been positive, with the main complaints/problems falling
into the following categories:
— Installation difficulties on some computers;

— Unavailability of a 32-bit version;
— Problems with screen resolution that causes the interface and corresponding fields to
appear unformatted;
— User error, due to not reading the manual;
— Initially a bug prevented the introduction of a new excitation function. This problem has
since been resolved;
— Some users found the graph in the main interface (Fig. 5.1) not useful and thought that
it occupied too much space. The graph is now optional.
The ERYA-profiling code will be available to the public on the same webpage as the ERYA-
bulk code, starting mid-April 2016. The page will also include a manual for the code.
220
Although not yet assessed by users, other tests have already been made, which were presented
at the 22nd International Conference on Ion Beam Analysis, June 14–19, 2015 - Opatija,
Croatia [5.8].
Taking into consideration the problems mentioned above, an effort is being made to
implement ERYA-bulk and ERYA-profiling in a different programming language and
platform interface, which will make it more suitable for the majority of users. Some
improvements, such as the possibility to use resonance strengths instead of cross-sections for
profiling, will be implemented.
[5.1] FONSECA, M., JESUS, A.P., LUÍS, H., MATEUS, R., CRUZ, J., GASQUES, L.,
Instrum. Methods Phys. Res. B 268 (2010) 1806–1908.
[5.2] MATEUS, R., JESUS, A.P., RIBEIRO, J.P., A code for quantitative analysis of light
elements in thick samples by PIGE, Nucl. Instrum. Methods Phys. Res. B 229 (2005)
302–308.
[5.3] LEVENBERG, K., A method for the solution of certain non-linear problems in least
squares, Quart. Appl. Math. 2 (1944) 164–168.
[5.4] MARQUARDT, D.W., An algorithm for least-squares estimation of nonlinear
parameters, J. Soc. Indust. Appl. Math. 11 (1963) 431–441.
[5.5] WANG, Y., NASTASI, M., (eds), Handbook of Modern Ion Beam Materials
Analysis, 2nd edn, Materials Research Society, Warrendale, Pennsylvania (2010).
[5.6] ROTONDI, A. AND MONTAGNA, P., Fast calculation of vavilov distribution,
[5.7] WILKINSON, D.H., Ionization energy loss by charged particles part I. The Landau
distribution, Nucl. Instrum. Methods Phys. Res. A 383 (1996) 513–515.
[5.8] BOGDANOVIĆ RADOVIĆ, I., JAKŠIĆ, M., KARLUŠIĆ, M., VIDOŠ, A., (Eds)
Book of Abstracts: 22nd International Conference on Ion Beam Analysis, Ruder
Boskovic Institute, Zagreb (2015).
221
6. CONCLUSIONS
Particle Induced Gamma Ray Emission (PIGE) is a well-established method of ion beam
analysis that is predominantly used for measurements performed relative to a well-known
standard sample. Due to the previous lack of many cross-section data, the full potential of this
technique could not be reached. The purpose of this CRP was to address the lack of a
comprehensive, up-to-date compilation of available PIGE cross-sections and yields for IBA
purposes, i.e. inelastic scattering and nuclear reactions, and the scarcity of pre-existing PIGE
data. The ultimate goal is to address the long-standing needs of the IBA community in
implementing a fully developed standardless PIGE analysis technique and the CRP was a
major step towards achieving this vision.
To meet these goals, a large number of new measurements needed to be performed, plus all
pre-existing data needed to be compiled and evaluated. In the end, these goals were
accomplished, which lead to an unprecedented amount of new data, plus an exhaustive
compilation of previous data, being uploaded to the Ion Beam Analysis Nuclear Data Library
(IBANDL).
Currently there are over 700 compiled data sets of gamma-producing reactions in IBANDL,
which is a huge increase from the approximately 70 sets before the CRP. IBANDL is now
fairly complete for most of the isotopes that are considered to be significant for PIGE analysis
of light projectiles. In order to successfully incorporate the new PIGE data into IBANDL, it
was necessary to modify the database. As part of the modification, it is now possible for
anyone to compare total isotropic cross-sections and differential cross-sections.
Since the new PIGE measurements were performed in ten different laboratories, it was
necessary to establish a common methodology for measuring differential cross-sections and
thick-target yields and to ensure consistent treatment of uncertainties. With the aid of a well-
defined inter-laboratory exercise, it was possible to identify the limitations and (lack of)
accuracy of the techniques. With this information, it could be decided which improvements
needed to be made in the measurement techniques and laboratory practices for a completely
standardless and accurate PIGE measurement.
Evaluated cross-section data are indispensable for a reliable and accurate implementation of
the standardless PIGE technique. During this CRP, theoretical evaluations of cross-sections
from gamma-ray producing reactions were performed, using an approach based on R-matrix
theory. Some of the recommended cross-sections have also become available to PIGE
practitioners.
In the absence of a complete set of evaluated cross-section data for PIGE analysis, alternative
ways of verifying and validating the measured cross-sections had to be employed. In this
CRP, special emphasis was placed on the comparison of measured cross-sections and thick-
target yields. Agreement between yields calculated from the measured cross-sections and
measured thick-target yields indicate a good understanding of the reaction and allows the
reliable analysis of samples of different thicknesses, without the use of individual standards.
This comparison technique was applied to all the data of gamma-ray producing reactions that
were measured and collected from the literature, i.e. reactions involving Li, Be, B, C, N, F,
Na, Mg, Al, Si, P, S, Ti and Cr. Although this project was extensive, additional measurements
of cross-sections and thick-target yields are needed for the following reactions in order to
properly validate all results:
222
14
— N, 28Si, 29Si, P, Cr, Ti, and deuteron-induced reactions in general, due to scarcity of
the existing data
7
— Li, 19F and 25Mg, due to discrepancies among the existing data.
In addition, for the needs of high-resolution depth profiling, which is one of the most
developed and advanced PIGE analysis techniques, IBANDL will need to be modified to
accommodate the energies, widths and strengths of narrow resonances.
Finally, a simulation code was developed and tested that combines cross-section and stopping
power data, over a given sample depth, to produce desired yields. This code was implemented
during the CRP. The ERYA code has been made available to the user community free of
charge in order to enhance the implementation of standardless PIGE analysis and enable its
comparative advantages in the analysis of in-depth homogeneous samples. The future release
of a simulation and analysis code for PIGE depth-profiling is expected to further promote the
use of standardless PIGE depth profiling, which would have significant benefits for the user
community at large.
223
Appendix
TABLE A.1. LIST OF REACTION CROSS-SECTIONS MEASURED DURING THE CRP (2011–2015)
Isotope Reaction γ-ray Energy Incident Energy Range Detection Angle Type of Data
[keV] [MeV]
7
Li (p,p′γ) 478 2–4, 2–10 130, 135 cross-sections + yields
7
Li (p,n′γ) 429 2–10 135 cross-sections + yields
9
Be (p,γ) 718 0.5–1.7 135 cross-sections + yields
10
B (p,p′γ) 718 1–3, 3–4, 3–5 0, 45, 55, 90, cross-sections + yields
165
10
B (p,α′γ) 429 1–3, 2–4, 3–4, 1–3.8 0, 45, 55, 90, yields
130, 165
11
B (p,p′γ) 2124 1–3, 3–4, 0, 45, 55, 90, cross-sections + yields
2.8–3.8 165
12
C (p,γ) 1.1–2.6 0, 55 cross-sections
12
C (d, p′γ ) 3089 0.6–2 55 cross-sections
14
N (p, p′γ ) 2313 4–7, 3.5–4 45, 55 cross-sections + yields
224
TABLE A.1. LIST OF REACTION CROSS-SECTIONS MEASURED DURING THE CRP (2011–2015) (cont.)
[keV] [MeV]
14
N (d, p′γ ) 1885 0.6–2 55, 90 cross-sections + yields
14
14
19
F (p, p′γ ) 110 2–4, 2–6 130, 135 cross-sections + yields
19
F (p, p′γ ) 197 2–4, 2–6 130, 135 cross-sections + yields
19
F (p,α′γ) 6000–7000 0.8–4, 2–6 130, 135 cross-sections + yields
23
Na (p, p′γ ) 440 2–4, 1.8–3, 3–4, 2–3, 0, 55, 90, 130, 135 cross-sections + yields
1–2.9,
23
Na (p, p′γ ) 1636 2–4, 1.8–3, 3–4, 2–3, 0, 55, 90, 130, 135 cross-sections + yields
(p, α′γ ) 1634 1–2.9,
24
Mg (p, p′γ ) 1369 1.9–3, 1–3, 2–6 0, 15, 55, 90, 135 cross-sections
25
Mg (p, p′γ ) 390 1.9–3, 1.5–2.4, 0, 15, 55, 90, 135, 130 cross-sections + yields
1–3, 2–6
25
Mg (p, p′γ ) 585 1.9–3, 1.5–2.4, 0, 15, 55, 90, 135, 130 cross-sections + yields
1–3, 2–6
25
Mg (p, p′γ ) 974 1.5–2.4, 1–3 90, 130 cross-sections + yields
225
[keV] [MeV]
26
Mg (p,γ) 1014 1–3 90 cross-sections
27
Al (p,p′γ) 844 1.5–3, 1–3, 2.5–5, 2.5–4, 0, 45, 55, 90, 135, 130, cross-sections + yields
2.5–3, 1.8–3, 1.5–3.8 165
27
Al (p,p′γ) 1014 1.5–3, 1–3, 2.5–5, 2.5–4, 0, 45, 55, 90, 135, 130, cross-sections + yields
2.5–3, 1.8–3, 1.5–3.8 165
27
Al (p,γ) 1779 1.5–4, 1–3, 2.5–5, 2–6, 0, 45, 55, 90, 135, 130, cross-sections + yields
2.5–3, 1.5–3.8 165
27
Al (p, α′γ ) 1369 1.5–4, 1–3, 2.5–5, 2–6, 0, 45, 55, 90, 135, 130, cross-sections + yields
1.5–3.8, 2.5–3 165
28
Si (p,p′γ) 1779 1–3, 4–7, 2.5–5 55, 90, 45 cross-sections + yields
28
Si (d, p′γ ) 1273 0.6–2 55 cross-sections
28
Si (d, p′γ ) 2028 0.6–2 55 cross-sections
29
Si (p,p′γ) 1274 1–3, 4–7, 2.5–5 55, 90, 45 cross-sections + yields
nat
Si (p,p′γ) 1274 2.5–4 45 yields
226
Isotope Reaction γ–ray Energy Incident Energy Range Detection Angle Type of Data
[keV] [MeV]
nat
Si (p, p′γ ) 1779 2.5–4 45 yields
31
P (p, p′γ ) 1266 1–3, 2–4 90, 130 cross-sections + yields
32
S (p, p′γ ) 2230 3–6 0, 15, 55, 90 cross-sections
32
S (d, p′γ ) 841 1–2 90 cross-sections + yields
48
Ti (p,γ) 62.3/90.6 1–3 90 cross-sections
nat
Ti (p,γ) 62.3/90.6 1–3 90 cross-sections
nat
Ti (p,nγ) 62.3/90.6 1.5–3 90 cross-sections
53
Cr (p,γ) 62.3/90.6 1–3 90 cross-sections
54
Cr (p,nγ) 62.3/90.6 1–3 90 cross-sections
227
LIST OF ABBREVIATIONS
ATOMKI ATOMKI Institute for Nuclear Research of the Hungarian Academy of

Sciences
CMAM-UAM Centro de Micro Análisis de Materiales, Universidad Autónoma de Madrid
CPSSR Coulomb-Repulsion Perturbed-Stationary State Relativistic Theory
CRP Coordinated Research Project
CTN/IST Campus Tecnologico Nuclear/Institut Superior Tecnico
DE Double Escape
EBS Elastic backscattering
ERDA Elastic Recoil Detection Analysis
ERYA Emitted Radiation Yield Analysis
FC Faraday Cup
FE Full Escape
FWHM Full width at half-maximum
Ge(HP) High Purity Germanium semiconductor detector
Ge(Li) Germanium semiconductor detector with Lithium doped crystals
GVM Generating VoltMeter
HVEE
IAEA International Atomic Energy Agency
IBA Ion beam analysis
IBANDL Ion Beam Analysis Data Library
KIPT Kharkov Institute of Physics and Technology
LABEC-INFN Laboratorio di Technice Nucleari per I Beni Culturali, Istituto Nazionale di
Fizica Nuclear, Firenze
LaBr Lanthanum Bromide scintillation radiation detector
MDL Minimum detection limit
MOCVD Metal-organic chemical vapor deposition
NaI Sodium Iodine detector
NDF General purpose code for analysis of Ion Beam Analysis Data
NDS Nuclear Data Section
NMR Nuclear Magnetic Resonance
NRA Nuclear reaction analysis
NSTRI Nuclear Science and Technology Research Institute
PIGE Particle-induced gamma-ray emission spectroscopy
PIPS Passivated Implanted Planar Silicon detector
PIXE Particle-induced X-ray emission spectroscopy
RBI Ruder Boskovic Institute
RBS Rutherford backscattering
RCM Research Coordination Meeting
REGe Reverse electrode coaxial Germanium detector
SDD Silicon drift detector
SIMNRA Microsoft Windows program for the simulation of backscattering spectra for
ion beam analysis with MeV ions
SSB Silicon Surface Barrier
229
LIST OF CONTRIBUTORS TO DRAFTING AND REVIEW
Becker, H-W. Ruhr Universität Bochum, Germany
Bogdanović Radović, I. Ruder Boskovic Institute, Croatia
Chiari, M. Istituto Nazionale di Fisica Nucleare, Italy
Goncharov, A. Kharkov Institute of Physics and Technology – National Science

Center Kharkov Institute of Physics & Technology, Ukraine
Kakuee, O. Nuclear Science & Technology Research Institute, Iran
Kiss, A.Z. Institute for Nuclear Research (ATOMKI) of the Hungarian

Academy of Sciences, Hungary
Lagoyannis, A. National Center of Scientific Research “Demokritos”, Greece
Pedro de Jesus, A. Centro de Fisica Nuclear, Portugal
Raisanen, J. University of Helsinki, Finland
Zucchiatti, A. Centro de Micro-Analisis de Materiales Universidad Autonóma

de Madrid, Spain
Dimitriou, P. Scientific Secretary
Abriola, D. Scientific Secretary
Research Coordination Meetings
Vienna, Austria: 16-20 May 2011
Vienna, Austria: 8-12 October 2012
Vienna, Austria: 7-11 April 2014
231
@ No. 25
ORDERING LOCALLY
In the following countries, IAEA priced publications may be purchased from the sources listed below or
from major local booksellers.
Orders for unpriced publications should be made directly to the IAEA. The contact details are given at
the end of this list.
CANADA
Renouf Publishing Co. Ltd
22-1010 Polytek Street, Ottawa, ON K1J 9J1, CANADA
Telephone: +1 613 745 2665  Fax: +1 643 745 7660
Email: order@renoufbooks.com  Web site: www.renoufbooks.com
Bernan / Rowman & Littlefield

15200 NBN Way, Blue Ridge Summit, PA 17214, USA
Tel: +1 800 462 6420 • Fax: +1 800 338 4550
Email: orders@rowman.com Web site: www.rowman.com/bernan
CZECH REPUBLIC
Suweco CZ, s.r.o.
Sestupná 153/11, 162 00 Prague 6, CZECH REPUBLIC
Telephone: +420 242 459 205  Fax: +420 284 821 646
Email: nakup@suweco.cz  Web site: www.suweco.cz
FRANCE
Form-Edit
5 rue Janssen, PO Box 25, 75921 Paris CEDEX, FRANCE
Telephone: +33 1 42 01 49 49  Fax: +33 1 42 01 90 90
Email: formedit@formedit.fr  Web site: www.form-edit.com
GERMANY
Goethe Buchhandlung Teubig GmbH
Schweitzer Fachinformationen
Willstätterstrasse 15, 40549 Düsseldorf, GERMANY
Telephone: +49 (0) 211 49 874 015  Fax: +49 (0) 211 49 874 28
Email: kundenbetreuung.goethe@schweitzer-online.de  Web site: www.goethebuch.de
INDIA
Allied Publishers
1st Floor, Dubash House, 15, J.N. Heredi Marg, Ballard Estate, Mumbai 400001, INDIA
Telephone: +91 22 4212 6930/31/69  Fax: +91 22 2261 7928
Email: alliedpl@vsnl.com  Web site: www.alliedpublishers.com
Bookwell
3/79 Nirankari, Delhi 110009, INDIA
Telephone: +91 11 2760 1283/4536
Email: bkwell@nde.vsnl.net.in  Web site: www.bookwellindia.com
ITALY
Libreria Scientifica “AEIOU”
Via Vincenzo Maria Coronelli 6, 20146 Milan, ITALY
Telephone: +39 02 48 95 45 52  Fax: +39 02 48 95 45 48
Email: info@libreriaaeiou.eu  Web site: www.libreriaaeiou.eu
JAPAN
Maruzen-Yushodo Co., Ltd
10-10 Yotsuyasakamachi, Shinjuku-ku, Tokyo 160-0002, JAPAN
Telephone: +81 3 4335 9312  Fax: +81 3 4335 9364
Email: bookimport@maruzen.co.jp  Web site: www.maruzen.co.jp
RUSSIAN FEDERATION
Scientific and Engineering Centre for Nuclear and Radiation Safety
107140, Moscow, Malaya Krasnoselskaya st. 2/8, bld. 5, RUSSIAN FEDERATION
Telephone: +7 499 264 00 03  Fax: +7 499 264 28 59
Email: secnrs@secnrs.ru  Web site: www.secnrs.ru
UNITED STATES OF AMERICA

Bernan / Rowman & Littlefield
15200 NBN Way, Blue Ridge Summit, PA 17214, USA
Tel: +1 800 462 6420 • Fax: +1 800 338 4550
Email: orders@rowman.com  Web site: www.rowman.com/bernan
Renouf Publishing Co. Ltd

812 Proctor Avenue, Ogdensburg, NY 13669-2205, USA
Telephone: +1 888 551 7470  Fax: +1 888 551 7471
Email: orders@renoufbooks.com  Web site: www.renoufbooks.com
Orders for both priced and unpriced publications may be addressed directly to:
Marketing and Sales Unit
17-31011
Vienna International Centre, PO Box 100, 1400 Vienna, Austria

Telephone: +43 1 2600 22529 or 22530 • Fax: +43 1 2600 29302 or +43 1 26007 22529
Email: sales.publications@iaea.org • Web site: www.iaea.org/books
IAEA-TECDOC-1822
IAEA-TECDOC-1822
IAEA TECDOC SERIES
Development of a Reference Database for Particle Induced Gamma Ray Emission (PIGE) Spectroscopy
IAEA-TECDOC-1822
Development of a Reference
Database for Particle Induced
Gamma Ray Emission
(PIGE) Spectroscopy

Vienna
ISBN 978–92–0–106317–5
ISSN 1011–4289
@
pc7209_TE-1822_cover.indd 1-3 2017-08-29 17:40:40

TE-1822 Web

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

TE-1822 Web

Uploaded by

Copyright:

Available Formats

IAEA-TECDOC-1822

International Atomic Energy Agency

AFGHANISTAN GEORGIA OMAN

INTERNATIONAL ATOMIC ENERGY AGENCY

Marketing and Sales Unit, Publishing Section

For further information on this publication, please contact:

Nuclear Data Section

IAEA Library Cataloguing in Publication Data

Names: International Atomic Energy Agency.

REFERENCES TO SECTION 1............................................................................................. 6

REFERENCES TO SECTION 2........................................................................................... 33

REFERENCES TO SECTION 4......................................................................................... 199

REFERENCES TO SECTION 5......................................................................................... 221

The IAEA CRP on ‘Development of a Reference Database for Particle-Induced Gamma-ray

1.3.1. Compilation of existing data

2) Identify discrepancies between measurements and which measurements were missing or

3) Compilation of retrieved data in IBANDL.

1.3.2. New measurements

1) New measurements of differential cross-sections and thick target yields of gamma-ray

Perform an inter-laboratory exercise to calibrate equipment and co-ordinate measurements:

1) Measurement of the excitation function of the 27Al(p,pγ1–0)27Al reaction, which is known

1.3.4. Evaluations of PIGE cross-sections

1.3.5. PIGE Analysis Code

[1.1] INTERNATIONAL ATOMIC ENERGY AGENCY, Development of a Reference

2.1. PIGE BULK ANALYSIS

2.1.1. PIGE bulk analysis

2.1.2. Production and acquisition of gamma-ray spectra

More details about a typical set-up are given in Section 2.3.

200 400 600 800 1000

1634 Na(p, αγ)20Ne

1634 Na(p, αγ)20Ne

1779* P(p, αγ)28Si

1779* P(p, αγ)28Si

* γ-rays emitted by more than one reaction

2.1.4. The standard-based method

This standard-based method has some drawbacks:

(i) It depends on the availability of standards, which should be as similar in composition to

2.1.5. The standardless method

where E0 is the incident proton energy.

2.2. PIGE DEPTH PROFILING WITH NUCLEAR RESONANCES

Depth profiling is performed by measuring the gamma-yield of the reaction as a function of

FIG. 2.3. Principle of depth profiling. Reproduced courtesy of Springer [2.44].

The Doppler broadening can be calculated by:

2.3. A TYPICAL MINIMAL PIGE FACILITY

Items a), b), and c) are elaborated in the following sub-sections.

2.3.1. Particle accelerator

TABLE 2.2. STATISTICS OF ACCELERATOR TYPES AND ACCELERATION TECHNOLOGY

2.3.2. Charge collection and measuring system

In external beams a relative charge monitor can be envisaged. It is a commonly-used device

2.3.3. Gamma-ray detector

2.3.4. Example of a minimal PIGE set-up

2.3.5. Example of an in-vacuum multi detector system

2.4. EXAMPLES OF APPLICATIONS IN NA BULK ANALYSIS

2.4.1. Identification of lapis-lazuli in paint layers

Measurements were performed at the KN3000 Van de Graaff accelerator of INFN in

2.4.2. Quantification of Na in ancient glasses

2.5. HYDROGEN DEPTH PROFILING

Hydrogen depth profiling by using a narrow resonance in a particle-induced gamma-ray

Pd Ti+O Ti Cr Si3N 4 SiO2

Accurate differential cross-section determination is dependent on the accuracy of recordings

3.1. ENERGY CALIBRATION OF THE ACCELERATOR

3.1.1. Narrow resonances